MedMNIST-Pneumonia.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# MedMINIST-Pneumonia\n",
    "\n",
    "Il seguente notebook illustra vari modelli per l'individuazione di soggetti positivi alla polmonite a partire da radiografie del torace in formato MNIST.\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 55,
   "metadata": {},
   "outputs": [],
   "source": [
    "import tensorflow as tf\n",
    "\n",
    "assert tf.__version__ >= \"2.16\"\n",
    "\n",
    "import keras_tuner as kt\n",
    "import numpy as np\n",
    "import pandas as pd\n",
    "import matplotlib.pyplot as plt\n",
    "import cv2\n",
    "import os\n",
    "\n",
    "from sklearn.ensemble import RandomForestClassifier\n",
    "from sklearn.linear_model import LogisticRegression\n",
    "from sklearn.metrics import (\n",
    "    make_scorer,\n",
    "    f1_score,\n",
    "    accuracy_score,\n",
    "    precision_recall_fscore_support,\n",
    "    ConfusionMatrixDisplay,\n",
    ")\n",
    "from sklearn.model_selection import train_test_split, StratifiedKFold\n",
    "from sklearn.utils.class_weight import compute_class_weight\n",
    "\n",
    "from imblearn.over_sampling import RandomOverSampler, SMOTE\n",
    "\n",
    "tf.keras.utils.set_random_seed(42)  # set all the seeds (Python, Keras, Tensorflow and NumPy)\n",
    "# tf.config.experimental.enable_op_determinism() # cant'use with some layers in CNN like MaxPool\n",
    "os.environ[\"TF_CUDNN_DETERMINISTIC\"] = \"1\"  # does not affect the model results"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "# plot functions\n",
    "\n",
    "plt.style.use(\"fivethirtyeight\")\n",
    "plt.rcParams[\"font.family\"] = \"sans-serif\"\n",
    "plt.rcParams[\"font.serif\"] = \"Ubuntu\"\n",
    "plt.rcParams[\"font.monospace\"] = \"Ubuntu Mono\"\n",
    "plt.rcParams[\"font.size\"] = 10\n",
    "plt.rcParams[\"axes.labelsize\"] = 10\n",
    "plt.rcParams[\"axes.labelweight\"] = \"bold\"\n",
    "plt.rcParams[\"axes.titlesize\"] = 10\n",
    "plt.rcParams[\"xtick.labelsize\"] = 8\n",
    "plt.rcParams[\"ytick.labelsize\"] = 8\n",
    "plt.rcParams[\"legend.fontsize\"] = 10\n",
    "plt.rcParams[\"figure.titlesize\"] = 12\n",
    "plt.rcParams[\"image.cmap\"] = \"jet\"\n",
    "plt.rcParams[\"image.interpolation\"] = \"none\"\n",
    "plt.rcParams[\"figure.figsize\"] = (16, 8)\n",
    "plt.rcParams[\"lines.linewidth\"] = 2\n",
    "plt.rcParams[\"lines.markersize\"] = 8\n",
    "\n",
    "\n",
    "def plot_measure(history_train, history_test, title=\"\", xticks=None):\n",
    "    plt.figure(figsize=(12, 8))\n",
    "    plt.plot(history_train, marker=\"o\", markersize=5, label=\"Train\")\n",
    "    plt.plot(history_test, marker=\"o\", markersize=5, label=\"Validation\")\n",
    "    plt.legend()\n",
    "    if xticks is None:\n",
    "        plt.gca().xaxis.set_major_locator(plt.NullLocator())\n",
    "    else:\n",
    "        ax = plt.subplot(111)\n",
    "        ax.set_xlim(xticks[0], xticks[-1])\n",
    "        plt.xticks(xticks)\n",
    "    plt.title(title)\n",
    "    plt.show()\n",
    "\n",
    "\n",
    "def plot_confusion_matrix(train_labels, predictions_train):\n",
    "    plt.figure(figsize=(12, 6))\n",
    "    disp = ConfusionMatrixDisplay.from_predictions(train_labels, predictions_train, normalize=\"true\", cmap=\"Blues\")\n",
    "    disp.figure_.suptitle(\"Confusion Matrix\")\n",
    "    plt.grid(None)\n",
    "    plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "# utilis function\n",
    "\n",
    "\n",
    "def print_accuracy_scores(train_labels, test_labels, predictions_train_model, predictions_test_model):\n",
    "    acc_train = accuracy_score(train_labels, predictions_train_model)\n",
    "    acc_test = accuracy_score(test_labels, predictions_test_model)\n",
    "    print(\"Training set accuracy:   {:.3f}\".format(acc_train))\n",
    "    print(\"Test set accuracy:       {:.3f}\".format(acc_test))\n",
    "\n",
    "\n",
    "def print_stats(labels, predictions, title):\n",
    "    meas = precision_recall_fscore_support(labels, predictions, average=None, zero_division=np.nan)\n",
    "    print(title)\n",
    "    for i in range(2):\n",
    "        print(\n",
    "            \"Class {0:d}: precision={1:5.3f}, recall={2:5.3f}, f1_score={3:5.3f}\".format(\n",
    "                i, meas[0][i], meas[1][i], meas[2][i]\n",
    "            )\n",
    "        )\n",
    "\n",
    "\n",
    "def train_validation_test_split(images, labels, test_val_size, verbose=False):\n",
    "    train_images, test_val_images, train_labels, test_val_labels = train_test_split(\n",
    "        images, labels, test_size=test_val_size, random_state=42\n",
    "    )\n",
    "    val_images, test_images, val_labels, test_labels = train_test_split(\n",
    "        test_val_images, test_val_labels, test_size=0.5, random_state=42\n",
    "    )\n",
    "\n",
    "    if verbose:\n",
    "        try:\n",
    "            print(\n",
    "                f\"Shape of training set of images:    {train_images.shape}\\t\\timages of size {train_images.shape[1]}x{train_images.shape[2]}\"\n",
    "            )\n",
    "            print(\n",
    "                f\"Shape of validation set of images:  {val_images.shape}\\t\\timages of size {val_images.shape[1]}x{val_images.shape[2]}\"\n",
    "            )\n",
    "            print(\n",
    "                f\"Shape of test set of images:        {test_images.shape}\\t\\timages of size {test_images.shape[1]}x{test_images.shape[2]}\"\n",
    "            )\n",
    "            nb_classes = np.unique(train_labels).shape[0]\n",
    "            print(f\"Number of classes: {nb_classes}\")\n",
    "        except IndexError:\n",
    "            print(f\"Shape of training set of images:    {train_images.shape}\")\n",
    "            print(f\"Shape of validation set of images:  {val_images.shape}\")\n",
    "            print(f\"Shape of test set of images:        {test_images.shape}\")\n",
    "            nb_classes = np.unique(train_labels).shape[0]\n",
    "            print(f\"Number of classes: {nb_classes}\")\n",
    "\n",
    "    return train_images, train_labels, val_images, val_labels, test_images, test_labels"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "The dataset is composed by 5232 images 28x28\n",
      "Negative examples:  1349\n",
      "Positive examples:  3883\n",
      "Imbalanced Dataset\n"
     ]
    }
   ],
   "source": [
    "# load dataset\n",
    "\n",
    "X = np.load(\"pneumonia_images.npy\")\n",
    "y = (np.load(\"pneumonia_labels.npy\")).ravel()\n",
    "\n",
    "if len(X) != len(y):\n",
    "    raise ValueError(\"Numbers of images and labels do not match\")\n",
    "\n",
    "print(f\"The dataset is composed by {X.shape[0]} images {X.shape[1]}x{X.shape[2]}\")\n",
    "\n",
    "bin_count = np.bincount(y)\n",
    "if len(bin_count) != 2:\n",
    "    raise ValueError(\"Labels must be 0 or 1\")\n",
    "\n",
    "print(\"Negative examples: \", bin_count[0])\n",
    "print(\"Positive examples: \", bin_count[1])\n",
    "\n",
    "if bin_count[0] != bin_count[1]:\n",
    "    print(\"Imbalanced Dataset\")\n",
    "else:\n",
    "    print(\"Balanced Dataset\")\n",
    "\n",
    "x_negative = next((i for i, x in enumerate(X) if y[i] == 0), None)\n",
    "x_positive = next((i for i, x in enumerate(X) if y[i] == 1), None)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Il dataset\n",
    "\n",
    "Il dataset viene fornite tramite due file .npy contenti le immagini e le rispettive etichette. Data la mancaza di documentazione si assume, banalmente, l'etichetta 0 come negativo e 1 come positivo.\n",
    "Dall'analisi del dataset è evidente uno sbilanciamento tra le classi:\n",
    "\n",
    "- 1349 elementi negativi;\n",
    "- 3883 elementi positivi.\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 840x360 with 21 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# print examples\n",
    "\n",
    "n_rows = 3\n",
    "n_cols = 7\n",
    "titles = [\"Negative\", \"Positive\"]\n",
    "plt.figure(figsize=(n_cols * 1.2, n_rows * 1.2))\n",
    "for row in range(n_rows):\n",
    "    for col in range(n_cols):\n",
    "        index = n_cols * row + col\n",
    "        plt.subplot(n_rows, n_cols, index + 1)\n",
    "        plt.imshow(X[index], cmap=\"binary\", interpolation=\"nearest\")\n",
    "        plt.axis(\"off\")\n",
    "        plt.title(titles[y[index]], fontsize=12)\n",
    "plt.subplots_adjust(wspace=0.2, hspace=0.5)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Istogrammi\n",
    "\n",
    "Il primo approccio applicato è quello degli istogrammi della tonalità di grigio. Da una breve visione delle immagine è evidente una tendenza al bianco delle radiografie dei soggetti malati. Questa è un'operazione di riduzione delle feature, dai 784 pixel di cui è composta l'immagine si passa al numero dei bin.\n",
    "Si è deciso di procedere all'analisi su istogrammi da 8, 16, 32 e 64 bin per i seguenti modelli:\n",
    "\n",
    "- Logistic Regression;\n",
    "- Random Forest.\n",
    "\n",
    "Per quando riguarda lo sbilanciamento tra le classi si è deciso di confrontare 3 soluzioni possibili:\n",
    "\n",
    "- aggiungere un peso agli elementi delle classi (Weighted);\n",
    "- Oversampling;\n",
    "- SMOTE.\n",
    "\n",
    "Il primo viene offerto da  _scikit-learn_  tramite il parametro _class_weight_; è sufficiente passare a questo parametro la stringa \"balanced\".\n",
    "\n",
    "L'oversamplig è una tecnica che consiste nel bilanciare le classi tramite la duplicazione degli elementi della classe minoritaria.\n",
    "\n",
    "SMOTE (Synthetic Minority Over-sampling Technique) è un metodo di oversampling che crea elementi sintetici tramite interpolazione dei vicini, nello spazio delle feature, degli elementi della classe minoritaria.\n",
    "\n",
    "Per Oversampling e SMOTE viene utilizzato il paccheto Imbalanced Learning.\n",
    "\n",
    "La ricerca degli iperparametri viene effetuata tramite Keras Tuner utilizzando SklearnTuner e il GridSearchOracle. L'80% del dataset viene destinato al Training Set e il rimanente al Test Set. Tutto il Trainig Set viene utilizzato per il tuning degli iperparametri che usa una cross validation, StrafiedKFold, con 5 fold. Come indicato dalla documentazione di Keras Tuner, una volta identificati i migliori iperparametri viene effetuato un nuovo training su tutto il Training Set.\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [],
   "source": [
    "def extract_histogram(\n",
    "    image,\n",
    "    color_space,\n",
    "    channels=(0, 1, 2),\n",
    "    bins=(8, 8, 8),\n",
    "    range=(0, 256, 0, 256, 0, 256),\n",
    "):\n",
    "    if color_space is None:\n",
    "        cvt_img = image\n",
    "    else:\n",
    "        cvt_img = cv2.cvtColor(image, color_space)\n",
    "    hist = cv2.calcHist([cvt_img], channels, None, bins, range)\n",
    "    cv2.normalize(hist, hist)\n",
    "    return hist.flatten()\n",
    "\n",
    "\n",
    "def get_histograms(dataset, color_space, channels, bins, channels_range):\n",
    "    histograms = []\n",
    "    for x in dataset:\n",
    "        hist = extract_histogram(x, color_space, channels, bins, channels_range)\n",
    "        histograms.append(hist)\n",
    "    return histograms\n",
    "\n",
    "\n",
    "def ranges(N, nb):\n",
    "    step = N / nb\n",
    "    return [\"{},{}\".format(round(step * i), round(step * (i + 1))) for i in range(nb)]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 1600x800 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 1600x800 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 1600x800 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 1600x800 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# negative\n",
    "hist8bins_negative = extract_histogram(X[x_negative], None, (0,), (8,), (0, 256))\n",
    "hist16bins_negative = extract_histogram(X[x_negative], None, (0,), (16,), (0, 256))\n",
    "hist32bins_negative = extract_histogram(X[x_negative], None, (0,), (32,), (0, 256))\n",
    "hist64bins_negative = extract_histogram(X[x_negative], None, (0,), (64,), (0, 256))\n",
    "\n",
    "# positive\n",
    "hist8bins__positive = extract_histogram(X[x_positive], None, (0,), (8,), (0, 256))\n",
    "hist16bins_positive = extract_histogram(X[x_positive], None, (0,), (16,), (0, 256))\n",
    "hist32bins_positive = extract_histogram(X[x_positive], None, (0,), (32,), (0, 256))\n",
    "hist64bins_positive = extract_histogram(X[x_positive], None, (0,), (64,), (0, 256))\n",
    "\n",
    "# 8 bins\n",
    "fig, axs = plt.subplots(1, 2, sharey=True, tight_layout=True)\n",
    "axs[0].bar(ranges(255, 8), hist8bins_negative)\n",
    "axs[1].bar(ranges(255, 8), hist8bins__positive, color=\"red\")\n",
    "plt.show()\n",
    "\n",
    "# 16 bins\n",
    "fig, axs = plt.subplots(1, 2, sharey=True, tight_layout=True)\n",
    "axs[0].bar(ranges(255, 16), hist16bins_negative)\n",
    "axs[1].bar(ranges(255, 16), hist16bins_positive, color=\"red\")\n",
    "plt.show()\n",
    "\n",
    "# 32 bins\n",
    "fig, axs = plt.subplots(1, 2, sharey=True, tight_layout=True)\n",
    "axs[0].bar(ranges(255, 32), hist32bins_negative)\n",
    "axs[1].bar(ranges(255, 32), hist32bins_positive, color=\"red\")\n",
    "plt.show()\n",
    "\n",
    "# 64 bins\n",
    "fig, axs = plt.subplots(1, 2, sharey=True, tight_layout=True)\n",
    "axs[0].bar(ranges(255, 64), hist64bins_negative)\n",
    "axs[1].bar(ranges(255, 64), hist64bins_positive, color=\"red\")\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [],
   "source": [
    "# histogram dataset\n",
    "\n",
    "bins_8 = np.array(get_histograms(X, None, (0,), (8,), (0, 256)))\n",
    "bins_16 = np.array(get_histograms(X, None, (0,), (16,), (0, 256)))\n",
    "bins_32 = np.array(get_histograms(X, None, (0,), (32,), (0, 256)))\n",
    "bins_64 = np.array(get_histograms(X, None, (0,), (64,), (0, 256)))\n",
    "\n",
    "ros = RandomOverSampler(random_state=42)\n",
    "bins_8_over, y_over = ros.fit_resample(bins_8, y)\n",
    "bins_16_over, y_over = ros.fit_resample(bins_16, y)\n",
    "bins_32_over, y_over = ros.fit_resample(bins_32, y)\n",
    "bins_64_over, y_over = ros.fit_resample(bins_64, y)\n",
    "\n",
    "sm = SMOTE(random_state=42)\n",
    "bins_8_smote, y_smote = sm.fit_resample(bins_8, y)\n",
    "bins_16_smote, y_smote = sm.fit_resample(bins_16, y)\n",
    "bins_32_smote, y_smote = sm.fit_resample(bins_32, y)\n",
    "bins_64_smote, y_smote = sm.fit_resample(bins_64, y)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Dato il problema medico si è posta l'attenzione non sull'accuracy, ma sulla recall della classe 1, i positivi, poichè è auspicabicale l'individuazione del malato a discapito di un sano che venga classificato erroneamente. Dato che il tuning degli iperparametri cercando la massimizzazione della recall della classe 1 portava spesso ad assegnare tutti gli elementi a quella classe si è prefirito utilizzare f1_score come metrica da massimizzare.\n",
    "\n",
    "Dai tre metodi di bilanciamento ci si aspettato diversi risultati nelle metriche precision, recall e F1. Per esempio, è noto che SMOTE tende a favorire la classe minoritaria, in questo caso è quella meno interessante per la nostra classificazione.\n",
    "\n",
    "Si ipotizza inoltre un miglioramento al crescre dei bin.\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [],
   "source": [
    "scorer = make_scorer(f1_score, pos_label=1, labels=[0, 1])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Logistic Regression\n",
    "\n",
    "Per la regressione logistica, la ricerca si è concentrata sulla penalità applicata al modello e sul suo peso.\n",
    "\n",
    "Utilizzando il solver **saga**, tra i tipi di regolarizzazione è possibile scegliere tra L1 (Lasso), L2 (Ridge) ed Elastic-Net. Elastic-Net permette di utlizzare L1 e L2 insieme pesando le 2 componenti tramite il parametro _l1_ratio_. Il valore di _l1_ratio_ è compreso tra 0 e 1, dove 0 corrisponde a L2 e 1 a L1.\n",
    "\n",
    "Gli iperparametri esaminati durante il tuning sono quindi 2:\n",
    "\n",
    "- **C**, l'inverso del peso della penalità;\n",
    "- **l1_ratio**.\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [],
   "source": [
    "def log_reg(project_name, data, labels, title, class_weight=None):\n",
    "\n",
    "    def build_logistic_regression(hp):\n",
    "        model = LogisticRegression(\n",
    "            n_jobs=-1,\n",
    "            random_state=42,\n",
    "            class_weight=class_weight,\n",
    "            C=hp.Choice(\"C\", [0.00001, 0.0001, 0.001, 0.01, 0.1, 1.0]),\n",
    "            solver=\"saga\",\n",
    "            penalty=\"elasticnet\",\n",
    "            l1_ratio=hp.Choice(\"l1_ratio\", [0.0, 0.1, 0.2, 0.3, 0.4, 0.5, 0.6, 0.7, 0.8, 0.9, 1.0]),\n",
    "            max_iter=200,\n",
    "        )\n",
    "        return model\n",
    "\n",
    "    tuner = kt.tuners.SklearnTuner(\n",
    "        oracle=kt.oracles.GridSearchOracle(objective=kt.Objective(\"score\", \"max\")),\n",
    "        scoring=scorer,\n",
    "        hypermodel=build_logistic_regression,\n",
    "        cv=StratifiedKFold(5),\n",
    "        project_name=project_name,\n",
    "    )\n",
    "\n",
    "    X_train, X_test, y_train, y_test = train_test_split(data, labels, test_size=0.2, random_state=42)\n",
    "\n",
    "    tuner.search(X_train, y_train)\n",
    "\n",
    "    best_hps = tuner.get_best_hyperparameters(num_trials=1)[0]\n",
    "\n",
    "    print(f\"\\n{title}\\n\")\n",
    "\n",
    "    print(\"Best C: \", best_hps.get(\"C\"))\n",
    "    print(\"Best l1_ratio: \", best_hps.get(\"l1_ratio\"))\n",
    "\n",
    "    print()\n",
    "\n",
    "    model0 = tuner.hypermodel.build(best_hps)\n",
    "    model0.fit(X_train, y_train)\n",
    "\n",
    "    predictions_test = model0.predict(X_test)\n",
    "    predictions_train = model0.predict(X_train)\n",
    "\n",
    "    print_accuracy_scores(y_train, y_test, predictions_train, predictions_test)\n",
    "    print_stats(y_train, predictions_train, \"Training set\")\n",
    "    print_stats(y_test, predictions_test, \"Test set\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Reloading Tuner from ./tuners/log_reg_hist_8bins_weighted/tuner0.json\n",
      "\n",
      "Logistic Regression 8 bins weighted\n",
      "\n",
      "Best C:  1.0\n",
      "Best l1_ratio:  1.0\n",
      "\n",
      "Training set accuracy:   0.810\n",
      "Test set accuracy:       0.796\n",
      "Training set\n",
      "Class 0: precision=0.594, recall=0.846, f1_score=0.698\n",
      "Class 1: precision=0.937, recall=0.797, f1_score=0.862\n",
      "Test set\n",
      "Class 0: precision=0.568, recall=0.808, f1_score=0.667\n",
      "Class 1: precision=0.924, recall=0.792, f1_score=0.853\n"
     ]
    }
   ],
   "source": [
    "log_reg(\n",
    "    \"tuners/log_reg_hist_8bins_weighted\",\n",
    "    bins_8,\n",
    "    y,\n",
    "    \"Logistic Regression 8 bins weighted\",\n",
    "    \"balanced\",\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Reloading Tuner from ./tuners/log_reg_hist_8bins_over/tuner0.json\n",
      "\n",
      "Logistic Regression 8 bins oversampling\n",
      "\n",
      "Best C:  1.0\n",
      "Best l1_ratio:  0.9\n",
      "\n",
      "Training set accuracy:   0.817\n",
      "Test set accuracy:       0.817\n",
      "Training set\n",
      "Class 0: precision=0.800, recall=0.843, f1_score=0.821\n",
      "Class 1: precision=0.837, recall=0.793, f1_score=0.814\n",
      "Test set\n",
      "Class 0: precision=0.815, recall=0.835, f1_score=0.825\n",
      "Class 1: precision=0.820, recall=0.798, f1_score=0.809\n"
     ]
    }
   ],
   "source": [
    "log_reg(\n",
    "    \"tuners/log_reg_hist_8bins_over\",\n",
    "    bins_8_over,\n",
    "    y_over,\n",
    "    \"Logistic Regression 8 bins oversampling\",\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Reloading Tuner from ./tuners/log_reg_hist_8bins_smote/tuner0.json\n",
      "\n",
      "Logistic Regression 8 bins SMOTE\n",
      "\n",
      "Best C:  1.0\n",
      "Best l1_ratio:  1.0\n",
      "\n",
      "Training set accuracy:   0.822\n",
      "Test set accuracy:       0.819\n",
      "Training set\n",
      "Class 0: precision=0.805, recall=0.845, f1_score=0.825\n",
      "Class 1: precision=0.840, recall=0.798, f1_score=0.819\n",
      "Test set\n",
      "Class 0: precision=0.814, recall=0.840, f1_score=0.827\n",
      "Class 1: precision=0.824, recall=0.795, f1_score=0.809\n"
     ]
    }
   ],
   "source": [
    "log_reg(\n",
    "    \"tuners/log_reg_hist_8bins_smote\",\n",
    "    bins_8_smote,\n",
    "    y_smote,\n",
    "    \"Logistic Regression 8 bins SMOTE\",\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Reloading Tuner from ./tuners/log_reg_hist_16bins_weighted/tuner0.json\n",
      "\n",
      "Logistic Regression 16 bins weighted\n",
      "\n",
      "Best C:  1.0\n",
      "Best l1_ratio:  1.0\n",
      "\n",
      "Training set accuracy:   0.832\n",
      "Test set accuracy:       0.818\n",
      "Training set\n",
      "Class 0: precision=0.626, recall=0.869, f1_score=0.728\n",
      "Class 1: precision=0.947, recall=0.819, f1_score=0.878\n",
      "Test set\n",
      "Class 0: precision=0.601, recall=0.834, f1_score=0.698\n",
      "Class 1: precision=0.935, recall=0.812, f1_score=0.869\n"
     ]
    }
   ],
   "source": [
    "log_reg(\n",
    "    \"tuners/log_reg_hist_16bins_weighted\",\n",
    "    bins_16,\n",
    "    y,\n",
    "    \"Logistic Regression 16 bins weighted\",\n",
    "    \"balanced\",\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Reloading Tuner from ./tuners/log_reg_hist_16bins_over/tuner0.json\n",
      "\n",
      "Logistic Regression 16 bins oversampling\n",
      "\n",
      "Best C:  1.0\n",
      "Best l1_ratio:  1.0\n",
      "\n",
      "Training set accuracy:   0.839\n",
      "Test set accuracy:       0.838\n",
      "Training set\n",
      "Class 0: precision=0.821, recall=0.862, f1_score=0.841\n",
      "Class 1: precision=0.857, recall=0.815, f1_score=0.836\n",
      "Test set\n",
      "Class 0: precision=0.835, recall=0.857, f1_score=0.846\n",
      "Class 1: precision=0.843, recall=0.819, f1_score=0.831\n"
     ]
    }
   ],
   "source": [
    "log_reg(\n",
    "    \"tuners/log_reg_hist_16bins_over\",\n",
    "    bins_16_over,\n",
    "    y_over,\n",
    "    \"Logistic Regression 16 bins oversampling\",\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Reloading Tuner from ./tuners/log_reg_hist_16bins_smote/tuner0.json\n",
      "\n",
      "Logistic Regression 16 bins SMOTE\n",
      "\n",
      "Best C:  1.0\n",
      "Best l1_ratio:  1.0\n",
      "\n",
      "Training set accuracy:   0.843\n",
      "Test set accuracy:       0.848\n",
      "Training set\n",
      "Class 0: precision=0.823, recall=0.870, f1_score=0.846\n",
      "Class 1: precision=0.865, recall=0.816, f1_score=0.840\n",
      "Test set\n",
      "Class 0: precision=0.839, recall=0.873, f1_score=0.856\n",
      "Class 1: precision=0.858, recall=0.822, f1_score=0.840\n"
     ]
    }
   ],
   "source": [
    "log_reg(\n",
    "    \"tuners/log_reg_hist_16bins_smote\",\n",
    "    bins_16_smote,\n",
    "    y_smote,\n",
    "    \"Logistic Regression 16 bins SMOTE\",\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Reloading Tuner from ./tuners/log_reg_hist_32bins_weighted/tuner0.json\n",
      "\n",
      "Logistic Regression 32 bins weighted\n",
      "\n",
      "Best C:  1.0\n",
      "Best l1_ratio:  1.0\n",
      "\n",
      "Training set accuracy:   0.836\n",
      "Test set accuracy:       0.819\n",
      "Training set\n",
      "Class 0: precision=0.633, recall=0.867, f1_score=0.732\n",
      "Class 1: precision=0.947, recall=0.825, f1_score=0.881\n",
      "Test set\n",
      "Class 0: precision=0.603, recall=0.838, f1_score=0.701\n",
      "Class 1: precision=0.937, recall=0.813, f1_score=0.871\n"
     ]
    }
   ],
   "source": [
    "log_reg(\n",
    "    \"tuners/log_reg_hist_32bins_weighted\",\n",
    "    bins_32,\n",
    "    y,\n",
    "    \"Logistic Regression 32 bins weighted\",\n",
    "    \"balanced\",\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Reloading Tuner from ./tuners/log_reg_hist_32bins_over/tuner0.json\n",
      "\n",
      "Logistic Regression 32 bins oversampling\n",
      "\n",
      "Best C:  1.0\n",
      "Best l1_ratio:  0.9\n",
      "\n",
      "Training set accuracy:   0.845\n",
      "Test set accuracy:       0.846\n",
      "Training set\n",
      "Class 0: precision=0.827, recall=0.870, f1_score=0.848\n",
      "Class 1: precision=0.865, recall=0.821, f1_score=0.843\n",
      "Test set\n",
      "Class 0: precision=0.842, recall=0.864, f1_score=0.853\n",
      "Class 1: precision=0.851, recall=0.827, f1_score=0.839\n"
     ]
    }
   ],
   "source": [
    "log_reg(\n",
    "    \"tuners/log_reg_hist_32bins_over\",\n",
    "    bins_32_over,\n",
    "    y_over,\n",
    "    \"Logistic Regression 32 bins oversampling\",\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Reloading Tuner from ./tuners/log_reg_hist_32bins_smote/tuner0.json\n",
      "\n",
      "Logistic Regression 32 bins SMOTE\n",
      "\n",
      "Best C:  1.0\n",
      "Best l1_ratio:  1.0\n",
      "\n",
      "Training set accuracy:   0.850\n",
      "Test set accuracy:       0.849\n",
      "Training set\n",
      "Class 0: precision=0.833, recall=0.873, f1_score=0.852\n",
      "Class 1: precision=0.869, recall=0.827, f1_score=0.848\n",
      "Test set\n",
      "Class 0: precision=0.845, recall=0.868, f1_score=0.856\n",
      "Class 1: precision=0.855, recall=0.830, f1_score=0.842\n"
     ]
    }
   ],
   "source": [
    "log_reg(\n",
    "    \"tuners/log_reg_hist_32bins_smote\",\n",
    "    bins_32_smote,\n",
    "    y_smote,\n",
    "    \"Logistic Regression 32 bins SMOTE\",\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Reloading Tuner from ./tuners/log_reg_hist_64bins_weighted/tuner0.json\n",
      "\n",
      "Logistic Regression 64 bins weighted\n",
      "\n",
      "Best C:  1.0\n",
      "Best l1_ratio:  1.0\n",
      "\n",
      "Training set accuracy:   0.836\n",
      "Test set accuracy:       0.816\n",
      "Training set\n",
      "Class 0: precision=0.633, recall=0.871, f1_score=0.733\n",
      "Class 1: precision=0.948, recall=0.823, f1_score=0.881\n",
      "Test set\n",
      "Class 0: precision=0.596, recall=0.842, f1_score=0.698\n",
      "Class 1: precision=0.938, recall=0.807, f1_score=0.867\n"
     ]
    }
   ],
   "source": [
    "log_reg(\n",
    "    \"tuners/log_reg_hist_64bins_weighted\",\n",
    "    bins_64,\n",
    "    y,\n",
    "    \"Logistic Regression 64 bins weighted\",\n",
    "    \"balanced\",\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Reloading Tuner from ./tuners/log_reg_hist_64bins_over/tuner0.json\n",
      "\n",
      "Logistic Regression 64 bins oversampling\n",
      "\n",
      "Best C:  1.0\n",
      "Best l1_ratio:  1.0\n",
      "\n",
      "Training set accuracy:   0.850\n",
      "Test set accuracy:       0.851\n",
      "Training set\n",
      "Class 0: precision=0.831, recall=0.874, f1_score=0.852\n",
      "Class 1: precision=0.870, recall=0.825, f1_score=0.847\n",
      "Test set\n",
      "Class 0: precision=0.840, recall=0.878, f1_score=0.859\n",
      "Class 1: precision=0.863, recall=0.822, f1_score=0.842\n"
     ]
    }
   ],
   "source": [
    "log_reg(\n",
    "    \"tuners/log_reg_hist_64bins_over\",\n",
    "    bins_64_over,\n",
    "    y_over,\n",
    "    \"Logistic Regression 64 bins oversampling\",\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Reloading Tuner from ./tuners/log_reg_hist_64bins_smote/tuner0.json\n",
      "\n",
      "Logistic Regression 64 bins SMOTE\n",
      "\n",
      "Best C:  1.0\n",
      "Best l1_ratio:  0.8\n",
      "\n",
      "Training set accuracy:   0.850\n",
      "Test set accuracy:       0.846\n",
      "Training set\n",
      "Class 0: precision=0.831, recall=0.876, f1_score=0.853\n",
      "Class 1: precision=0.871, recall=0.825, f1_score=0.847\n",
      "Test set\n",
      "Class 0: precision=0.838, recall=0.870, f1_score=0.854\n",
      "Class 1: precision=0.856, recall=0.820, f1_score=0.838\n"
     ]
    }
   ],
   "source": [
    "log_reg(\n",
    "    \"tuners/log_reg_hist_64bins_smote\",\n",
    "    bins_64_smote,\n",
    "    y_smote,\n",
    "    \"Logistic Regression 64 bins SMOTE\",\n",
    ")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "In tutti i casi è stato selezionato durante il tuning la panalità L1 con il costo minore.\n",
    "\n",
    "Come ipotizzato, l'aumento del numero di feature ha portato a un miglioramento di tutte le metriche considerate, ma fino a 32 bin.\n",
    "\n",
    "Per quanto riguarda i tre metodi di bilanciamento, il risultato migliore sulla metrica utilizzata per il tuning è stato ottenuto applicando i pesi alle classi. Tuttavia, l'accuracy è risultata migliore con SMOTE, seguita dall'Oversampling, e infine dalla metodo dei pesi (Weighted).\n",
    "\n",
    "È importante notare che, nel caso del metodo Weighted, c'è una significativa differenza tra la classe 0 e la classe 1 in termini di Precision e Recall. In tutti i casi, si osserva uno sbilanciamento notevole a favore della classe maggioritaria. Al contrario, con SMOTE si verifica l'effetto opposto come previsto.\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Random Forest\n",
    "\n",
    "Il numero di parametri offerti da **RandomForestClassifier** è molto più ampio rispetto a quello del modello precedente. Per questo motivo, la ricerca degli iperparametri ottimali è stata limitata ai seguenti tre:\n",
    "\n",
    "- **n_estimators**: il numero di alberi nella foresta.\n",
    "- **max_depth**: la profondità massima degli alberi.\n",
    "- **ccp_alpha**, parametro usato per il Minimal Cost-Complexity Pruning, una tecnica per rimuovere i nodi non foglia sotto una certa soglia di complessità.\n",
    "\n",
    "Inoltre non viene permessa la formazioni di nodi foglia con un solo elemento.\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [],
   "source": [
    "def rand_forest(\n",
    "    project_name,\n",
    "    data,\n",
    "    labels,\n",
    "    title,\n",
    "    class_weight=None,\n",
    "    n_estimators=[100, 250],\n",
    "    max_depth=[15, 20, 25],\n",
    "    ccp_alpha=[0.0005, 0.001, 0.0015, 0.002],\n",
    "):\n",
    "\n",
    "    def build_random_forest(hp):\n",
    "        model = RandomForestClassifier(\n",
    "            n_estimators=hp.Choice(\"n_estimators\", n_estimators),\n",
    "            max_depth=hp.Choice(\"max_depth\", max_depth),\n",
    "            min_samples_leaf=2,\n",
    "            ccp_alpha=hp.Choice(\"ccp_alpha\", ccp_alpha),\n",
    "            class_weight=class_weight,\n",
    "            n_jobs=-1,\n",
    "            random_state=42,\n",
    "        )\n",
    "\n",
    "        return model\n",
    "\n",
    "    tuner = kt.tuners.SklearnTuner(\n",
    "        oracle=kt.oracles.GridSearchOracle(objective=kt.Objective(\"score\", \"max\")),\n",
    "        scoring=scorer,\n",
    "        hypermodel=build_random_forest,\n",
    "        cv=StratifiedKFold(5),\n",
    "        overwrite=False,\n",
    "        project_name=project_name,\n",
    "    )\n",
    "\n",
    "    X_train, X_test, y_train, y_test = train_test_split(data, labels, test_size=0.2, random_state=42)\n",
    "\n",
    "    tuner.search(X_train, y_train)\n",
    "\n",
    "    best_hps = tuner.get_best_hyperparameters(num_trials=1)[0]\n",
    "\n",
    "    print(f\"\\n{title}\\n\")\n",
    "\n",
    "    print(\"Best n_estimators:       \", best_hps.get(\"n_estimators\"))\n",
    "    print(\"Best max_depth:          \", best_hps.get(\"max_depth\"))\n",
    "    print(\"Best ccp_alpha:          \", best_hps.get(\"ccp_alpha\"))\n",
    "\n",
    "    print()\n",
    "\n",
    "    model0 = tuner.hypermodel.build(best_hps)\n",
    "    model0.fit(X_train, y_train)\n",
    "\n",
    "    predictions_test = model0.predict(X_test)\n",
    "    predictions_train = model0.predict(X_train)\n",
    "\n",
    "    print_accuracy_scores(y_train, y_test, predictions_train, predictions_test)\n",
    "    print_stats(y_train, predictions_train, \"Training set\")\n",
    "    print_stats(y_test, predictions_test, \"Test set\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Reloading Tuner from ./tuners/rand_forest_hist_8bins_weighted/tuner0.json\n",
      "\n",
      "Random Forest 8 bins weighted\n",
      "\n",
      "Best n_estimators:        100\n",
      "Best max_depth:           25\n",
      "Best ccp_alpha:           0.0005\n",
      "\n",
      "Training set accuracy:   0.954\n",
      "Test set accuracy:       0.874\n",
      "Training set\n",
      "Class 0: precision=0.880, recall=0.951, f1_score=0.914\n",
      "Class 1: precision=0.982, recall=0.955, f1_score=0.968\n",
      "Test set\n",
      "Class 0: precision=0.755, recall=0.743, f1_score=0.749\n",
      "Class 1: precision=0.913, recall=0.918, f1_score=0.916\n"
     ]
    }
   ],
   "source": [
    "rand_forest(\n",
    "    \"tuners/rand_forest_hist_8bins_weighted\",\n",
    "    bins_8,\n",
    "    y,\n",
    "    \"Random Forest 8 bins weighted\",\n",
    "    \"balanced\",\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Reloading Tuner from ./tuners/rand_forest_hist_8bins_over/tuner0.json\n",
      "\n",
      "Random Forest 8 bins oversampling\n",
      "\n",
      "Best n_estimators:        250\n",
      "Best max_depth:           25\n",
      "Best ccp_alpha:           0.0005\n",
      "\n",
      "Training set accuracy:   0.952\n",
      "Test set accuracy:       0.912\n",
      "Training set\n",
      "Class 0: precision=0.941, recall=0.964, f1_score=0.953\n",
      "Class 1: precision=0.964, recall=0.941, f1_score=0.952\n",
      "Test set\n",
      "Class 0: precision=0.905, recall=0.926, f1_score=0.916\n",
      "Class 1: precision=0.920, recall=0.896, f1_score=0.908\n"
     ]
    }
   ],
   "source": [
    "rand_forest(\n",
    "    \"tuners/rand_forest_hist_8bins_over\",\n",
    "    bins_8_over,\n",
    "    y_over,\n",
    "    \"Random Forest 8 bins oversampling\",\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Reloading Tuner from ./tuners/rand_forest_hist_8bins_smote/tuner0.json\n",
      "\n",
      "Random Forest 8 bins SMOTE\n",
      "\n",
      "Best n_estimators:        250\n",
      "Best max_depth:           20\n",
      "Best ccp_alpha:           0.0005\n",
      "\n",
      "Training set accuracy:   0.942\n",
      "Test set accuracy:       0.894\n",
      "Training set\n",
      "Class 0: precision=0.936, recall=0.948, f1_score=0.942\n",
      "Class 1: precision=0.948, recall=0.936, f1_score=0.942\n",
      "Test set\n",
      "Class 0: precision=0.907, recall=0.887, f1_score=0.897\n",
      "Class 1: precision=0.882, recall=0.903, f1_score=0.892\n"
     ]
    }
   ],
   "source": [
    "rand_forest(\n",
    "    \"tuners/rand_forest_hist_8bins_smote\",\n",
    "    bins_8_smote,\n",
    "    y_smote,\n",
    "    \"Random Forest 8 bins SMOTE\",\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Reloading Tuner from ./tuners/rand_forest_hist_16bins_weighted/tuner0.json\n",
      "\n",
      "Random Forest 16 bins weighted\n",
      "\n",
      "Best n_estimators:        250\n",
      "Best max_depth:           15\n",
      "Best ccp_alpha:           0.0005\n",
      "\n",
      "Training set accuracy:   0.960\n",
      "Test set accuracy:       0.897\n",
      "Training set\n",
      "Class 0: precision=0.888, recall=0.969, f1_score=0.926\n",
      "Class 1: precision=0.989, recall=0.957, f1_score=0.973\n",
      "Test set\n",
      "Class 0: precision=0.785, recall=0.815, f1_score=0.800\n",
      "Class 1: precision=0.937, recall=0.925, f1_score=0.931\n"
     ]
    }
   ],
   "source": [
    "rand_forest(\n",
    "    \"tuners/rand_forest_hist_16bins_weighted\",\n",
    "    bins_16,\n",
    "    y,\n",
    "    \"Random Forest 16 bins weighted\",\n",
    "    \"balanced\",\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Reloading Tuner from ./tuners/rand_forest_hist_16bins_over/tuner0.json\n",
      "\n",
      "Random Forest 16 bins oversampling\n",
      "\n",
      "Best n_estimators:        250\n",
      "Best max_depth:           25\n",
      "Best ccp_alpha:           0.0005\n",
      "\n",
      "Training set accuracy:   0.967\n",
      "Test set accuracy:       0.940\n",
      "Training set\n",
      "Class 0: precision=0.955, recall=0.979, f1_score=0.967\n",
      "Class 1: precision=0.978, recall=0.955, f1_score=0.966\n",
      "Test set\n",
      "Class 0: precision=0.923, recall=0.965, f1_score=0.943\n",
      "Class 1: precision=0.961, recall=0.914, f1_score=0.937\n"
     ]
    }
   ],
   "source": [
    "rand_forest(\n",
    "    \"tuners/rand_forest_hist_16bins_over\",\n",
    "    bins_16_over,\n",
    "    y_over,\n",
    "    \"Random Forest 16 bins oversampling\",\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Reloading Tuner from ./tuners/rand_forest_hist_16bins_smote/tuner0.json\n",
      "\n",
      "Random Forest 16 bins SMOTE\n",
      "\n",
      "Best n_estimators:        250\n",
      "Best max_depth:           20\n",
      "Best ccp_alpha:           0.0005\n",
      "\n",
      "Training set accuracy:   0.958\n",
      "Test set accuracy:       0.918\n",
      "Training set\n",
      "Class 0: precision=0.953, recall=0.964, f1_score=0.958\n",
      "Class 1: precision=0.964, recall=0.953, f1_score=0.959\n",
      "Test set\n",
      "Class 0: precision=0.919, recall=0.921, f1_score=0.920\n",
      "Class 1: precision=0.916, recall=0.914, f1_score=0.915\n"
     ]
    }
   ],
   "source": [
    "rand_forest(\n",
    "    \"tuners/rand_forest_hist_16bins_smote\",\n",
    "    bins_16_smote,\n",
    "    y_smote,\n",
    "    \"Random Forest 16 bins SMOTE\",\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Reloading Tuner from ./tuners/rand_forest_hist_32bins_weighted/tuner0.json\n",
      "\n",
      "Random Forest 32 bins weighted\n",
      "\n",
      "Best n_estimators:        100\n",
      "Best max_depth:           20\n",
      "Best ccp_alpha:           0.0005\n",
      "\n",
      "Training set accuracy:   0.971\n",
      "Test set accuracy:       0.893\n",
      "Training set\n",
      "Class 0: precision=0.917, recall=0.976, f1_score=0.945\n",
      "Class 1: precision=0.991, recall=0.969, f1_score=0.980\n",
      "Test set\n",
      "Class 0: precision=0.778, recall=0.808, f1_score=0.793\n",
      "Class 1: precision=0.934, recall=0.922, f1_score=0.928\n"
     ]
    }
   ],
   "source": [
    "rand_forest(\n",
    "    \"tuners/rand_forest_hist_32bins_weighted\",\n",
    "    bins_32,\n",
    "    y,\n",
    "    \"Random Forest 32 bins weighted\",\n",
    "    \"balanced\",\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Reloading Tuner from ./tuners/rand_forest_hist_32bins_over/tuner0.json\n",
      "\n",
      "Random Forest 32 bins oversampling\n",
      "\n",
      "Best n_estimators:        100\n",
      "Best max_depth:           25\n",
      "Best ccp_alpha:           0.0005\n",
      "\n",
      "Training set accuracy:   0.972\n",
      "Test set accuracy:       0.932\n",
      "Training set\n",
      "Class 0: precision=0.960, recall=0.984, f1_score=0.972\n",
      "Class 1: precision=0.984, recall=0.960, f1_score=0.972\n",
      "Test set\n",
      "Class 0: precision=0.915, recall=0.958, f1_score=0.936\n",
      "Class 1: precision=0.952, recall=0.906, f1_score=0.928\n"
     ]
    }
   ],
   "source": [
    "rand_forest(\n",
    "    \"tuners/rand_forest_hist_32bins_over\",\n",
    "    bins_32_over,\n",
    "    y_over,\n",
    "    \"Random Forest 32 bins oversampling\",\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Reloading Tuner from ./tuners/rand_forest_hist_32bins_smote/tuner0.json\n",
      "\n",
      "Random Forest 32 bins SMOTE\n",
      "\n",
      "Best n_estimators:        100\n",
      "Best max_depth:           25\n",
      "Best ccp_alpha:           0.0005\n",
      "\n",
      "Training set accuracy:   0.968\n",
      "Test set accuracy:       0.928\n",
      "Training set\n",
      "Class 0: precision=0.961, recall=0.974, f1_score=0.968\n",
      "Class 1: precision=0.974, recall=0.961, f1_score=0.968\n",
      "Test set\n",
      "Class 0: precision=0.924, recall=0.938, f1_score=0.931\n",
      "Class 1: precision=0.932, recall=0.918, f1_score=0.925\n"
     ]
    }
   ],
   "source": [
    "rand_forest(\n",
    "    \"tuners/rand_forest_hist_32bins_smote\",\n",
    "    bins_32_smote,\n",
    "    y_smote,\n",
    "    \"Random Forest 32 bins SMOTE\",\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Reloading Tuner from ./tuners/rand_forest_hist_64bins_weighted/tuner0.json\n",
      "\n",
      "Random Forest 64 bins weighted\n",
      "\n",
      "Best n_estimators:        250\n",
      "Best max_depth:           20\n",
      "Best ccp_alpha:           0.0005\n",
      "\n",
      "Training set accuracy:   0.978\n",
      "Test set accuracy:       0.885\n",
      "Training set\n",
      "Class 0: precision=0.934, recall=0.983, f1_score=0.958\n",
      "Class 1: precision=0.994, recall=0.976, f1_score=0.985\n",
      "Test set\n",
      "Class 0: precision=0.774, recall=0.774, f1_score=0.774\n",
      "Class 1: precision=0.923, recall=0.923, f1_score=0.923\n"
     ]
    }
   ],
   "source": [
    "rand_forest(\n",
    "    \"tuners/rand_forest_hist_64bins_weighted\",\n",
    "    bins_64,\n",
    "    y,\n",
    "    \"Random Forest 64 bins weighted\",\n",
    "    \"balanced\",\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Reloading Tuner from ./tuners/rand_forest_hist_64bins_over/tuner0.json\n",
      "\n",
      "Random Forest 64 bins oversampling\n",
      "\n",
      "Best n_estimators:        250\n",
      "Best max_depth:           20\n",
      "Best ccp_alpha:           0.0005\n",
      "\n",
      "Training set accuracy:   0.978\n",
      "Test set accuracy:       0.936\n",
      "Training set\n",
      "Class 0: precision=0.968, recall=0.989, f1_score=0.978\n",
      "Class 1: precision=0.989, recall=0.967, f1_score=0.978\n",
      "Test set\n",
      "Class 0: precision=0.914, recall=0.968, f1_score=0.940\n",
      "Class 1: precision=0.963, recall=0.903, f1_score=0.932\n"
     ]
    }
   ],
   "source": [
    "rand_forest(\n",
    "    \"tuners/rand_forest_hist_64bins_over\",\n",
    "    bins_64_over,\n",
    "    y_over,\n",
    "    \"Random Forest 64 bins oversampling\",\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Reloading Tuner from ./tuners/rand_forest_hist_64bins_smote/tuner0.json\n",
      "\n",
      "Random Forest 64 bins SMOTE\n",
      "\n",
      "Best n_estimators:        250\n",
      "Best max_depth:           25\n",
      "Best ccp_alpha:           0.0005\n",
      "\n",
      "Training set accuracy:   0.973\n",
      "Test set accuracy:       0.920\n",
      "Training set\n",
      "Class 0: precision=0.968, recall=0.978, f1_score=0.973\n",
      "Class 1: precision=0.978, recall=0.968, f1_score=0.973\n",
      "Test set\n",
      "Class 0: precision=0.913, recall=0.934, f1_score=0.924\n",
      "Class 1: precision=0.928, recall=0.906, f1_score=0.917\n"
     ]
    }
   ],
   "source": [
    "rand_forest(\n",
    "    \"tuners/rand_forest_hist_64bins_smote\",\n",
    "    bins_64_smote,\n",
    "    y_smote,\n",
    "    \"Random Forest 64 bins SMOTE\",\n",
    ")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## CNN\n",
    "\n",
    "L'ultimo modello che si vuole presentare è una rete convoluzionale che utilizza per far fronte allo sbilanciamento del dataset dei pesi per le classi.\n",
    "\n",
    "Verrano prensentate 2 varianti della stessa rete che diffirisconi per il layer finale:\n",
    "\n",
    "- Dense con un nodo e come funzione di attivazione la sigmoide (cnn0);\n",
    "- Dense con 2 nodi e come funzione di attivazione la softmax (cnn1).\n",
    "\n",
    "Questa variazione è necessaria per effettuare il tuning su F1 score poiché viene richiesto un set di label onehot.\n",
    "\n",
    "A causa di questa scelta verranno effettuati 3 tuning:\n",
    "\n",
    "- cnn0 con obiettivo la massimizzazione dell'accuracy\n",
    "- cnn1 con obiettivo la massimizzazione dell'accuracy\n",
    "- cnn1 con obiettivo la massimizzazione di F1\n",
    "\n",
    "Riguardo F1 non è disponibile una metrica simile a quella utilizza in precedenza. Infatti f1_score restituisce un array contenente il valore per classe che non può essere utilizzato in fase di tuning. Come soluzione si è scelto di utilizzare f1_score(\"weighted) che calcola i valori di F1 per entrambe le classi e ne restiusce una media pesata.\n",
    "\n",
    "In questo caso il dataset viene diviso come segue:\n",
    "\n",
    "- 70% Training Set;\n",
    "- 15% Validation Set;\n",
    "- 15% Test Set.\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Shape of training set of images:    (3662, 28, 28)\t\timages of size 28x28\n",
      "Shape of validation set of images:  (785, 28, 28)\t\timages of size 28x28\n",
      "Shape of test set of images:        (785, 28, 28)\t\timages of size 28x28\n",
      "Number of classes: 2\n"
     ]
    }
   ],
   "source": [
    "y = np.load(\"pneumonia_labels.npy\")\n",
    "\n",
    "X_train, y_train, X_valid, y_valid, X_test, y_test = train_validation_test_split(X, y, 0.3, 1)\n",
    "\n",
    "class_labels = np.unique(y_train)\n",
    "class_weights = compute_class_weight(class_weight=\"balanced\", classes=class_labels, y=y_train.ravel())\n",
    "class_weights = dict(zip(class_labels, class_weights))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "La CNN che si è testata consta dei seguenti layer:\n",
    "\n",
    "- 2 layer di prepocessing, GaussianNoise e Rescaling;\n",
    "- 3 layer convulozionali, composti da Conv2D, MaxPool2D;\n",
    "- un layer Flatten, per passare dalla matrice al vettore,\n",
    "- 2 layer Dense ognuno seguito da un layer Dropout,\n",
    "- il layer di output."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 57,
   "metadata": {},
   "outputs": [],
   "source": [
    "filters = [[12, 32, 48], [24, 48, 96]]\n",
    "dense_units = [[128, 256], [256, 512]]\n",
    "rate = [0.1, 0.2, 0.3, 0.4, 0.5]\n",
    "activations = [\"relu\", \"leaky_relu\"]\n",
    "reg = [0.0, 0.1, 0.01, 0.001]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "\n",
    "Gli iperparametri testati sono i seguenti:\n",
    "\n",
    "- i filtri da applicare alle convoluzioni e il numero di nodi dei layer Dense;\n",
    "- la funzione di attivazione da applicare;\n",
    "- il dropout rate dei layer Dropout (uguale per tutti i layer);\n",
    "- la regolarizzazione del kernel dei layer Dense (uguale per tutti i layer).\n",
    "\n",
    "Per la regolarizzazione si è scelto utilizzare L1 e L2 testando anche la loro assenza.\n",
    "\n",
    "I pesi vengono inizializzati con HeNormal e le immagini vengono preprocessate inserendo un rumore gaussiano con stddev pari a 0.2 ."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 56,
   "metadata": {},
   "outputs": [],
   "source": [
    "binary_accuracy = tf.keras.metrics.BinaryAccuracy()\n",
    "obj_binary_accuracy = kt.Objective(\"val_binary_accuracy\", \"max\")\n",
    "\n",
    "\n",
    "def cnn0_builder(hp):\n",
    "    index = hp.Int(\"index\", min_value=0, max_value=1, step=1)\n",
    "    dropout_rate = hp.Choice(\"dropout_rate\", rate)\n",
    "    l1 = hp.Choice(\"L1\", reg)\n",
    "    l2 = hp.Choice(\"L2\", reg)\n",
    "    activation = hp.Choice(\"activations\", activations)\n",
    "\n",
    "    init = tf.keras.initializers.HeNormal()\n",
    "\n",
    "    model0 = tf.keras.Sequential()\n",
    "\n",
    "    model0.add(tf.keras.layers.InputLayer((28, 28, 1)))\n",
    "    model0.add(tf.keras.layers.Rescaling(1.0 / 255))\n",
    "    model0.add(tf.keras.layers.GaussianNoise(0.2))\n",
    "\n",
    "    model0.add(tf.keras.layers.Conv2D(filters[index][0], 3, padding=\"same\", activation=activation))\n",
    "    model0.add(tf.keras.layers.MaxPool2D())\n",
    "\n",
    "    model0.add(tf.keras.layers.Conv2D(filters[index][1], 3, padding=\"same\", activation=activation))\n",
    "    model0.add(tf.keras.layers.MaxPool2D())\n",
    "\n",
    "    model0.add(tf.keras.layers.Conv2D(filters[index][2], 3, padding=\"same\", activation=activation))\n",
    "    model0.add(tf.keras.layers.MaxPool2D())\n",
    "\n",
    "    model0.add(tf.keras.layers.Flatten())\n",
    "\n",
    "    model0.add(\n",
    "        tf.keras.layers.Dense(\n",
    "            dense_units[index][1],\n",
    "            activation,\n",
    "            kernel_initializer=init,\n",
    "            kernel_regularizer=tf.keras.regularizers.L1L2(l1, l2),\n",
    "        ),\n",
    "    )\n",
    "    model0.add(tf.keras.layers.Dropout(dropout_rate))\n",
    "\n",
    "    model0.add(\n",
    "        tf.keras.layers.Dense(\n",
    "            dense_units[index][0],\n",
    "            activation,\n",
    "            kernel_initializer=init,\n",
    "            kernel_regularizer=tf.keras.regularizers.L1L2(l1, l2),\n",
    "        )\n",
    "    )\n",
    "    model0.add(tf.keras.layers.Dropout(dropout_rate))\n",
    "\n",
    "    model0.add(\n",
    "        tf.keras.layers.Dense(\n",
    "            1,\n",
    "            \"sigmoid\",\n",
    "            kernel_initializer=init,\n",
    "            kernel_regularizer=tf.keras.regularizers.L1L2(l1, l2),\n",
    "        )\n",
    "    )\n",
    "\n",
    "    model0.compile(optimizer=\"adam\", loss=\"binary_crossentropy\", metrics=[binary_accuracy])\n",
    "\n",
    "    return model0"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 39,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Reloading Tuner from ./tuners/cnn0_weighted/tuner0.json\n"
     ]
    }
   ],
   "source": [
    "early_stop = tf.keras.callbacks.EarlyStopping(monitor=\"loss\", mode=\"min\", patience=3)\n",
    "\n",
    "tuner = kt.Hyperband(cnn0_builder, obj_binary_accuracy, seed=42, project_name=\"tuners/cnn0_weighted\")\n",
    "tuner.search(X_train, y_train, validation_data=(X_valid, y_valid), callbacks=[early_stop], class_weight=class_weights)\n",
    "best_hps = tuner.get_best_hyperparameters(num_trials=1)[0]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 40,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Best Filters:  [24, 48, 96]\n",
      "Best Dense Units:  [256, 512]\n",
      "Best Activation Function:  leaky_relu\n",
      "Best Dropout Rate:  0.2\n"
     ]
    }
   ],
   "source": [
    "print(\"Best Filters: \", filters[best_hps.get(\"index\")])\n",
    "print(\"Best Dense Units: \", dense_units[best_hps.get(\"index\")])\n",
    "print(\"Best Activation Function: \", best_hps.get(\"activations\"))\n",
    "\n",
    "if best_hps.get(\"L1\") > 0:\n",
    "    print(\"Best L1 Value: \", best_hps.get(\"L1\"))\n",
    "\n",
    "if best_hps.get(\"L2\") > 0:\n",
    "    print(\"Best L2 Value: \", best_hps.get(\"L2\"))\n",
    "\n",
    "print(\"Best Dropout Rate: \", best_hps.get(\"dropout_rate\"))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 41,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<pre style=\"white-space:pre;overflow-x:auto;line-height:normal;font-family:Menlo,'DejaVu Sans Mono',consolas,'Courier New',monospace\"><span style=\"font-weight: bold\">Model: \"sequential\"</span>\n",
       "</pre>\n"
      ],
      "text/plain": [
       "\u001b[1mModel: \"sequential\"\u001b[0m\n"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre style=\"white-space:pre;overflow-x:auto;line-height:normal;font-family:Menlo,'DejaVu Sans Mono',consolas,'Courier New',monospace\">┏━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━┓\n",
       "┃<span style=\"font-weight: bold\"> Layer (type)                    </span>┃<span style=\"font-weight: bold\"> Output Shape           </span>┃<span style=\"font-weight: bold\">       Param # </span>┃\n",
       "┡━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━┩\n",
       "│ rescaling (<span style=\"color: #0087ff; text-decoration-color: #0087ff\">Rescaling</span>)           │ (<span style=\"color: #00d7ff; text-decoration-color: #00d7ff\">None</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">28</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">28</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">1</span>)      │             <span style=\"color: #00af00; text-decoration-color: #00af00\">0</span> │\n",
       "├─────────────────────────────────┼────────────────────────┼───────────────┤\n",
       "│ gaussian_noise (<span style=\"color: #0087ff; text-decoration-color: #0087ff\">GaussianNoise</span>)  │ (<span style=\"color: #00d7ff; text-decoration-color: #00d7ff\">None</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">28</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">28</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">1</span>)      │             <span style=\"color: #00af00; text-decoration-color: #00af00\">0</span> │\n",
       "├─────────────────────────────────┼────────────────────────┼───────────────┤\n",
       "│ conv2d (<span style=\"color: #0087ff; text-decoration-color: #0087ff\">Conv2D</span>)                 │ (<span style=\"color: #00d7ff; text-decoration-color: #00d7ff\">None</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">28</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">28</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">24</span>)     │           <span style=\"color: #00af00; text-decoration-color: #00af00\">240</span> │\n",
       "├─────────────────────────────────┼────────────────────────┼───────────────┤\n",
       "│ max_pooling2d (<span style=\"color: #0087ff; text-decoration-color: #0087ff\">MaxPooling2D</span>)    │ (<span style=\"color: #00d7ff; text-decoration-color: #00d7ff\">None</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">14</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">14</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">24</span>)     │             <span style=\"color: #00af00; text-decoration-color: #00af00\">0</span> │\n",
       "├─────────────────────────────────┼────────────────────────┼───────────────┤\n",
       "│ conv2d_1 (<span style=\"color: #0087ff; text-decoration-color: #0087ff\">Conv2D</span>)               │ (<span style=\"color: #00d7ff; text-decoration-color: #00d7ff\">None</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">14</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">14</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">48</span>)     │        <span style=\"color: #00af00; text-decoration-color: #00af00\">10,416</span> │\n",
       "├─────────────────────────────────┼────────────────────────┼───────────────┤\n",
       "│ max_pooling2d_1 (<span style=\"color: #0087ff; text-decoration-color: #0087ff\">MaxPooling2D</span>)  │ (<span style=\"color: #00d7ff; text-decoration-color: #00d7ff\">None</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">7</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">7</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">48</span>)       │             <span style=\"color: #00af00; text-decoration-color: #00af00\">0</span> │\n",
       "├─────────────────────────────────┼────────────────────────┼───────────────┤\n",
       "│ conv2d_2 (<span style=\"color: #0087ff; text-decoration-color: #0087ff\">Conv2D</span>)               │ (<span style=\"color: #00d7ff; text-decoration-color: #00d7ff\">None</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">7</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">7</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">96</span>)       │        <span style=\"color: #00af00; text-decoration-color: #00af00\">41,568</span> │\n",
       "├─────────────────────────────────┼────────────────────────┼───────────────┤\n",
       "│ max_pooling2d_2 (<span style=\"color: #0087ff; text-decoration-color: #0087ff\">MaxPooling2D</span>)  │ (<span style=\"color: #00d7ff; text-decoration-color: #00d7ff\">None</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">3</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">3</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">96</span>)       │             <span style=\"color: #00af00; text-decoration-color: #00af00\">0</span> │\n",
       "├─────────────────────────────────┼────────────────────────┼───────────────┤\n",
       "│ flatten (<span style=\"color: #0087ff; text-decoration-color: #0087ff\">Flatten</span>)               │ (<span style=\"color: #00d7ff; text-decoration-color: #00d7ff\">None</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">864</span>)            │             <span style=\"color: #00af00; text-decoration-color: #00af00\">0</span> │\n",
       "├─────────────────────────────────┼────────────────────────┼───────────────┤\n",
       "│ dense (<span style=\"color: #0087ff; text-decoration-color: #0087ff\">Dense</span>)                   │ (<span style=\"color: #00d7ff; text-decoration-color: #00d7ff\">None</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">512</span>)            │       <span style=\"color: #00af00; text-decoration-color: #00af00\">442,880</span> │\n",
       "├─────────────────────────────────┼────────────────────────┼───────────────┤\n",
       "│ dropout (<span style=\"color: #0087ff; text-decoration-color: #0087ff\">Dropout</span>)               │ (<span style=\"color: #00d7ff; text-decoration-color: #00d7ff\">None</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">512</span>)            │             <span style=\"color: #00af00; text-decoration-color: #00af00\">0</span> │\n",
       "├─────────────────────────────────┼────────────────────────┼───────────────┤\n",
       "│ dense_1 (<span style=\"color: #0087ff; text-decoration-color: #0087ff\">Dense</span>)                 │ (<span style=\"color: #00d7ff; text-decoration-color: #00d7ff\">None</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">256</span>)            │       <span style=\"color: #00af00; text-decoration-color: #00af00\">131,328</span> │\n",
       "├─────────────────────────────────┼────────────────────────┼───────────────┤\n",
       "│ dropout_1 (<span style=\"color: #0087ff; text-decoration-color: #0087ff\">Dropout</span>)             │ (<span style=\"color: #00d7ff; text-decoration-color: #00d7ff\">None</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">256</span>)            │             <span style=\"color: #00af00; text-decoration-color: #00af00\">0</span> │\n",
       "├─────────────────────────────────┼────────────────────────┼───────────────┤\n",
       "│ dense_2 (<span style=\"color: #0087ff; text-decoration-color: #0087ff\">Dense</span>)                 │ (<span style=\"color: #00d7ff; text-decoration-color: #00d7ff\">None</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">1</span>)              │           <span style=\"color: #00af00; text-decoration-color: #00af00\">257</span> │\n",
       "└─────────────────────────────────┴────────────────────────┴───────────────┘\n",
       "</pre>\n"
      ],
      "text/plain": [
       "┏━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━┓\n",
       "┃\u001b[1m \u001b[0m\u001b[1mLayer (type)                   \u001b[0m\u001b[1m \u001b[0m┃\u001b[1m \u001b[0m\u001b[1mOutput Shape          \u001b[0m\u001b[1m \u001b[0m┃\u001b[1m \u001b[0m\u001b[1m      Param #\u001b[0m\u001b[1m \u001b[0m┃\n",
       "┡━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━┩\n",
       "│ rescaling (\u001b[38;5;33mRescaling\u001b[0m)           │ (\u001b[38;5;45mNone\u001b[0m, \u001b[38;5;34m28\u001b[0m, \u001b[38;5;34m28\u001b[0m, \u001b[38;5;34m1\u001b[0m)      │             \u001b[38;5;34m0\u001b[0m │\n",
       "├─────────────────────────────────┼────────────────────────┼───────────────┤\n",
       "│ gaussian_noise (\u001b[38;5;33mGaussianNoise\u001b[0m)  │ (\u001b[38;5;45mNone\u001b[0m, \u001b[38;5;34m28\u001b[0m, \u001b[38;5;34m28\u001b[0m, \u001b[38;5;34m1\u001b[0m)      │             \u001b[38;5;34m0\u001b[0m │\n",
       "├─────────────────────────────────┼────────────────────────┼───────────────┤\n",
       "│ conv2d (\u001b[38;5;33mConv2D\u001b[0m)                 │ (\u001b[38;5;45mNone\u001b[0m, \u001b[38;5;34m28\u001b[0m, \u001b[38;5;34m28\u001b[0m, \u001b[38;5;34m24\u001b[0m)     │           \u001b[38;5;34m240\u001b[0m │\n",
       "├─────────────────────────────────┼────────────────────────┼───────────────┤\n",
       "│ max_pooling2d (\u001b[38;5;33mMaxPooling2D\u001b[0m)    │ (\u001b[38;5;45mNone\u001b[0m, \u001b[38;5;34m14\u001b[0m, \u001b[38;5;34m14\u001b[0m, \u001b[38;5;34m24\u001b[0m)     │             \u001b[38;5;34m0\u001b[0m │\n",
       "├─────────────────────────────────┼────────────────────────┼───────────────┤\n",
       "│ conv2d_1 (\u001b[38;5;33mConv2D\u001b[0m)               │ (\u001b[38;5;45mNone\u001b[0m, \u001b[38;5;34m14\u001b[0m, \u001b[38;5;34m14\u001b[0m, \u001b[38;5;34m48\u001b[0m)     │        \u001b[38;5;34m10,416\u001b[0m │\n",
       "├─────────────────────────────────┼────────────────────────┼───────────────┤\n",
       "│ max_pooling2d_1 (\u001b[38;5;33mMaxPooling2D\u001b[0m)  │ (\u001b[38;5;45mNone\u001b[0m, \u001b[38;5;34m7\u001b[0m, \u001b[38;5;34m7\u001b[0m, \u001b[38;5;34m48\u001b[0m)       │             \u001b[38;5;34m0\u001b[0m │\n",
       "├─────────────────────────────────┼────────────────────────┼───────────────┤\n",
       "│ conv2d_2 (\u001b[38;5;33mConv2D\u001b[0m)               │ (\u001b[38;5;45mNone\u001b[0m, \u001b[38;5;34m7\u001b[0m, \u001b[38;5;34m7\u001b[0m, \u001b[38;5;34m96\u001b[0m)       │        \u001b[38;5;34m41,568\u001b[0m │\n",
       "├─────────────────────────────────┼────────────────────────┼───────────────┤\n",
       "│ max_pooling2d_2 (\u001b[38;5;33mMaxPooling2D\u001b[0m)  │ (\u001b[38;5;45mNone\u001b[0m, \u001b[38;5;34m3\u001b[0m, \u001b[38;5;34m3\u001b[0m, \u001b[38;5;34m96\u001b[0m)       │             \u001b[38;5;34m0\u001b[0m │\n",
       "├─────────────────────────────────┼────────────────────────┼───────────────┤\n",
       "│ flatten (\u001b[38;5;33mFlatten\u001b[0m)               │ (\u001b[38;5;45mNone\u001b[0m, \u001b[38;5;34m864\u001b[0m)            │             \u001b[38;5;34m0\u001b[0m │\n",
       "├─────────────────────────────────┼────────────────────────┼───────────────┤\n",
       "│ dense (\u001b[38;5;33mDense\u001b[0m)                   │ (\u001b[38;5;45mNone\u001b[0m, \u001b[38;5;34m512\u001b[0m)            │       \u001b[38;5;34m442,880\u001b[0m │\n",
       "├─────────────────────────────────┼────────────────────────┼───────────────┤\n",
       "│ dropout (\u001b[38;5;33mDropout\u001b[0m)               │ (\u001b[38;5;45mNone\u001b[0m, \u001b[38;5;34m512\u001b[0m)            │             \u001b[38;5;34m0\u001b[0m │\n",
       "├─────────────────────────────────┼────────────────────────┼───────────────┤\n",
       "│ dense_1 (\u001b[38;5;33mDense\u001b[0m)                 │ (\u001b[38;5;45mNone\u001b[0m, \u001b[38;5;34m256\u001b[0m)            │       \u001b[38;5;34m131,328\u001b[0m │\n",
       "├─────────────────────────────────┼────────────────────────┼───────────────┤\n",
       "│ dropout_1 (\u001b[38;5;33mDropout\u001b[0m)             │ (\u001b[38;5;45mNone\u001b[0m, \u001b[38;5;34m256\u001b[0m)            │             \u001b[38;5;34m0\u001b[0m │\n",
       "├─────────────────────────────────┼────────────────────────┼───────────────┤\n",
       "│ dense_2 (\u001b[38;5;33mDense\u001b[0m)                 │ (\u001b[38;5;45mNone\u001b[0m, \u001b[38;5;34m1\u001b[0m)              │           \u001b[38;5;34m257\u001b[0m │\n",
       "└─────────────────────────────────┴────────────────────────┴───────────────┘\n"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre style=\"white-space:pre;overflow-x:auto;line-height:normal;font-family:Menlo,'DejaVu Sans Mono',consolas,'Courier New',monospace\"><span style=\"font-weight: bold\"> Total params: </span><span style=\"color: #00af00; text-decoration-color: #00af00\">626,689</span> (2.39 MB)\n",
       "</pre>\n"
      ],
      "text/plain": [
       "\u001b[1m Total params: \u001b[0m\u001b[38;5;34m626,689\u001b[0m (2.39 MB)\n"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre style=\"white-space:pre;overflow-x:auto;line-height:normal;font-family:Menlo,'DejaVu Sans Mono',consolas,'Courier New',monospace\"><span style=\"font-weight: bold\"> Trainable params: </span><span style=\"color: #00af00; text-decoration-color: #00af00\">626,689</span> (2.39 MB)\n",
       "</pre>\n"
      ],
      "text/plain": [
       "\u001b[1m Trainable params: \u001b[0m\u001b[38;5;34m626,689\u001b[0m (2.39 MB)\n"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre style=\"white-space:pre;overflow-x:auto;line-height:normal;font-family:Menlo,'DejaVu Sans Mono',consolas,'Courier New',monospace\"><span style=\"font-weight: bold\"> Non-trainable params: </span><span style=\"color: #00af00; text-decoration-color: #00af00\">0</span> (0.00 B)\n",
       "</pre>\n"
      ],
      "text/plain": [
       "\u001b[1m Non-trainable params: \u001b[0m\u001b[38;5;34m0\u001b[0m (0.00 B)\n"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "model0 = tuner.hypermodel.build(best_hps)\n",
    "model0.summary()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 42,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/100\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "I0000 00:00:1727279341.918655    8944 service.cc:146] XLA service 0x7f46bc063ff0 initialized for platform CUDA (this does not guarantee that XLA will be used). Devices:\n",
      "I0000 00:00:1727279341.918676    8944 service.cc:154]   StreamExecutor device (0): NVIDIA GeForce MX450, Compute Capability 7.5\n",
      "2024-09-25 17:49:01.978168: I tensorflow/compiler/mlir/tensorflow/utils/dump_mlir_util.cc:268] disabling MLIR crash reproducer, set env var `MLIR_CRASH_REPRODUCER_DIRECTORY` to enable.\n",
      "2024-09-25 17:49:02.163386: I external/local_xla/xla/stream_executor/cuda/cuda_dnn.cc:531] Loaded cuDNN version 90201\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\u001b[1m 29/115\u001b[0m \u001b[32m━━━━━\u001b[0m\u001b[37m━━━━━━━━━━━━━━━\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - binary_accuracy: 0.6072 - loss: 0.6905"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "I0000 00:00:1727279344.945446    8944 device_compiler.h:188] Compiled cluster using XLA!  This line is logged at most once for the lifetime of the process.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\u001b[1m115/115\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 47ms/step - binary_accuracy: 0.6665 - loss: 0.5913\n",
      "Epoch 1: val_binary_accuracy improved from -inf to 0.89809, saving model to ./best_model/cnn0_weighted.weights.h5\n",
      "\u001b[1m115/115\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m11s\u001b[0m 55ms/step - binary_accuracy: 0.6674 - loss: 0.5900 - val_binary_accuracy: 0.8981 - val_loss: 0.2360\n",
      "Epoch 2/100\n",
      "\u001b[1m104/115\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m━━\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - binary_accuracy: 0.8952 - loss: 0.2435\n",
      "Epoch 2: val_binary_accuracy did not improve from 0.89809\n",
      "\u001b[1m115/115\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m1s\u001b[0m 5ms/step - binary_accuracy: 0.8952 - loss: 0.2432 - val_binary_accuracy: 0.8981 - val_loss: 0.2362\n",
      "Epoch 3/100\n",
      "\u001b[1m104/115\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m━━\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - binary_accuracy: 0.9112 - loss: 0.2007\n",
      "Epoch 3: val_binary_accuracy improved from 0.89809 to 0.93503, saving model to ./best_model/cnn0_weighted.weights.h5\n",
      "\u001b[1m115/115\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m1s\u001b[0m 5ms/step - binary_accuracy: 0.9111 - loss: 0.2018 - val_binary_accuracy: 0.9350 - val_loss: 0.1720\n",
      "Epoch 4/100\n",
      "\u001b[1m107/115\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m━━\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - binary_accuracy: 0.9210 - loss: 0.1882\n",
      "Epoch 4: val_binary_accuracy did not improve from 0.93503\n",
      "\u001b[1m115/115\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m1s\u001b[0m 5ms/step - binary_accuracy: 0.9207 - loss: 0.1889 - val_binary_accuracy: 0.9236 - val_loss: 0.1959\n",
      "Epoch 5/100\n",
      "\u001b[1m105/115\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m━━\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - binary_accuracy: 0.9243 - loss: 0.1890\n",
      "Epoch 5: val_binary_accuracy improved from 0.93503 to 0.94777, saving model to ./best_model/cnn0_weighted.weights.h5\n",
      "\u001b[1m115/115\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m1s\u001b[0m 6ms/step - binary_accuracy: 0.9239 - loss: 0.1901 - val_binary_accuracy: 0.9478 - val_loss: 0.1429\n",
      "Epoch 6/100\n",
      "\u001b[1m104/115\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m━━\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - binary_accuracy: 0.9290 - loss: 0.1678\n",
      "Epoch 6: val_binary_accuracy improved from 0.94777 to 0.94904, saving model to ./best_model/cnn0_weighted.weights.h5\n",
      "\u001b[1m115/115\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m1s\u001b[0m 5ms/step - binary_accuracy: 0.9286 - loss: 0.1691 - val_binary_accuracy: 0.9490 - val_loss: 0.1404\n",
      "Epoch 7/100\n",
      "\u001b[1m105/115\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m━━\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - binary_accuracy: 0.9252 - loss: 0.1705\n",
      "Epoch 7: val_binary_accuracy did not improve from 0.94904\n",
      "\u001b[1m115/115\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m1s\u001b[0m 5ms/step - binary_accuracy: 0.9252 - loss: 0.1706 - val_binary_accuracy: 0.9452 - val_loss: 0.1425\n",
      "Epoch 8/100\n",
      "\u001b[1m115/115\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - binary_accuracy: 0.9330 - loss: 0.1614\n",
      "Epoch 8: val_binary_accuracy improved from 0.94904 to 0.95032, saving model to ./best_model/cnn0_weighted.weights.h5\n",
      "\u001b[1m115/115\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m1s\u001b[0m 5ms/step - binary_accuracy: 0.9329 - loss: 0.1615 - val_binary_accuracy: 0.9503 - val_loss: 0.1283\n",
      "Epoch 9/100\n",
      "\u001b[1m104/115\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m━━\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - binary_accuracy: 0.9367 - loss: 0.1509\n",
      "Epoch 9: val_binary_accuracy improved from 0.95032 to 0.95541, saving model to ./best_model/cnn0_weighted.weights.h5\n",
      "\u001b[1m115/115\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m1s\u001b[0m 5ms/step - binary_accuracy: 0.9363 - loss: 0.1522 - val_binary_accuracy: 0.9554 - val_loss: 0.1267\n",
      "Epoch 10/100\n",
      "\u001b[1m113/115\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m━\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - binary_accuracy: 0.9314 - loss: 0.1702\n",
      "Epoch 10: val_binary_accuracy did not improve from 0.95541\n",
      "\u001b[1m115/115\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m1s\u001b[0m 5ms/step - binary_accuracy: 0.9313 - loss: 0.1703 - val_binary_accuracy: 0.9529 - val_loss: 0.1364\n",
      "Epoch 11/100\n",
      "\u001b[1m104/115\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m━━\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - binary_accuracy: 0.9342 - loss: 0.1422\n",
      "Epoch 11: val_binary_accuracy did not improve from 0.95541\n",
      "\u001b[1m115/115\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m1s\u001b[0m 5ms/step - binary_accuracy: 0.9341 - loss: 0.1429 - val_binary_accuracy: 0.9516 - val_loss: 0.1295\n",
      "Epoch 12/100\n",
      "\u001b[1m105/115\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m━━\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - binary_accuracy: 0.9350 - loss: 0.1550\n",
      "Epoch 12: val_binary_accuracy did not improve from 0.95541\n",
      "\u001b[1m115/115\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m1s\u001b[0m 5ms/step - binary_accuracy: 0.9352 - loss: 0.1545 - val_binary_accuracy: 0.9465 - val_loss: 0.1366\n",
      "Epoch 13/100\n",
      "\u001b[1m107/115\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m━━\u001b[0m \u001b[1m0s\u001b[0m 4ms/step - binary_accuracy: 0.9414 - loss: 0.1334\n",
      "Epoch 13: val_binary_accuracy did not improve from 0.95541\n",
      "\u001b[1m115/115\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m1s\u001b[0m 5ms/step - binary_accuracy: 0.9409 - loss: 0.1347 - val_binary_accuracy: 0.9503 - val_loss: 0.1307\n",
      "Epoch 14/100\n",
      "\u001b[1m106/115\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m━━\u001b[0m \u001b[1m0s\u001b[0m 4ms/step - binary_accuracy: 0.9403 - loss: 0.1432\n",
      "Epoch 14: val_binary_accuracy improved from 0.95541 to 0.95796, saving model to ./best_model/cnn0_weighted.weights.h5\n",
      "\u001b[1m115/115\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m1s\u001b[0m 5ms/step - binary_accuracy: 0.9402 - loss: 0.1433 - val_binary_accuracy: 0.9580 - val_loss: 0.1253\n",
      "Epoch 15/100\n",
      "\u001b[1m109/115\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m━━\u001b[0m \u001b[1m0s\u001b[0m 4ms/step - binary_accuracy: 0.9424 - loss: 0.1380\n",
      "Epoch 15: val_binary_accuracy improved from 0.95796 to 0.95924, saving model to ./best_model/cnn0_weighted.weights.h5\n",
      "\u001b[1m115/115\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m1s\u001b[0m 5ms/step - binary_accuracy: 0.9422 - loss: 0.1382 - val_binary_accuracy: 0.9592 - val_loss: 0.1217\n",
      "Epoch 16/100\n",
      "\u001b[1m109/115\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m━━\u001b[0m \u001b[1m0s\u001b[0m 4ms/step - binary_accuracy: 0.9410 - loss: 0.1478\n",
      "Epoch 16: val_binary_accuracy did not improve from 0.95924\n",
      "\u001b[1m115/115\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m1s\u001b[0m 5ms/step - binary_accuracy: 0.9409 - loss: 0.1478 - val_binary_accuracy: 0.9554 - val_loss: 0.1302\n",
      "Epoch 17/100\n",
      "\u001b[1m108/115\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m━━\u001b[0m \u001b[1m0s\u001b[0m 4ms/step - binary_accuracy: 0.9438 - loss: 0.1322\n",
      "Epoch 17: val_binary_accuracy did not improve from 0.95924\n",
      "\u001b[1m115/115\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m1s\u001b[0m 5ms/step - binary_accuracy: 0.9434 - loss: 0.1331 - val_binary_accuracy: 0.9516 - val_loss: 0.1363\n",
      "Epoch 18/100\n",
      "\u001b[1m105/115\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m━━\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - binary_accuracy: 0.9397 - loss: 0.1407\n",
      "Epoch 18: val_binary_accuracy did not improve from 0.95924\n",
      "\u001b[1m115/115\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m1s\u001b[0m 5ms/step - binary_accuracy: 0.9392 - loss: 0.1416 - val_binary_accuracy: 0.9465 - val_loss: 0.1335\n"
     ]
    }
   ],
   "source": [
    "model_file = \"./best_model/cnn0_weighted.weights.h5\"\n",
    "checkpoint = tf.keras.callbacks.ModelCheckpoint(model_file, \"val_binary_accuracy\", 1, True, True, \"max\")\n",
    "history_model0 = model0.fit(\n",
    "    X_train,\n",
    "    y_train,\n",
    "    epochs=100,\n",
    "    validation_data=(X_valid, y_valid),\n",
    "    callbacks=[early_stop, checkpoint],\n",
    "    class_weight=class_weights,\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 43,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 1200x800 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 1200x800 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\u001b[1m115/115\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 2ms/step\n",
      "\u001b[1m25/25\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "<Figure size 1200x600 with 0 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 1600x800 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "<Figure size 1200x600 with 0 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAzgAAAMOCAYAAADSr6MyAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjkuMiwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy8hTgPZAAAACXBIWXMAAA9hAAAPYQGoP6dpAABIv0lEQVR4nO3deZiWZdk/8O8NiIaCw74NW6aCimgoLrhiYrjmFm6FmW+LiZqpZf3SyrfUFpcyKVOzXHFNS9PcUtwTFVfSTFBBEUFcA1nm9wc6byPbDAzPjDefD8fzHsw91/Nc1z1Hh+98Oc/nfIpZs2bVBAAAoARaNPUBAAAAGouAAwAAlIaAAwAAlIaAAwAAlIaAAwAAlIaAAwAAlIaAAwAAlIaAAwAAlIaAAwAAlIaAQ+k8+eSTOeKII7Lxxhuna9eu6dmzZ7bbbrucffbZeeONN1bq3hMmTMiuu+6a3r17p6qqKueee26j71FVVZVTTz210V93WS699NJUVVWlqqoq48aNW+T7NTU12XTTTVNVVZXddtttufY4//zzc+mllzboOePGjVvimQCAVU+rpj4ANKY//OEP+da3vpV11103Rx11VNZff/3Mmzcvjz76aC688MI89NBDDf4FuiGOPPLIvPfee7ngggtSVVWV3r17N/oet956a3r06NHor1tfbdu2zcUXX5xtt922zvV77rknL7zwQtq2bbvcr33BBRekQ4cOOfjgg+v9nEGDBuXWW2/N+uuvv9z7AgDlIeBQGg899FCOPfbY7Ljjjrn00kuz+uqr135vxx13zJFHHpnbbrttpZ7hmWeeyahRo7LzzjuvtD0233zzlfba9bH33nvnqquuyltvvZV27drVXr/44oszZMiQvP322xU5x9y5c1MURdq1a9fkPxMAoPnQokZp/OIXv0hRFDnrrLPqhJsPtW7dOrvuumvt1wsWLMjZZ5+dzTffPF26dMmnPvWpfPWrX82UKVPqPG+33XbLVlttlUceeSQjRoxI9+7dM2jQoJx55plZsGBBkv9r35o3b15t9aaqqipJcuqpp9b+/b99+JzJkyfXXrvrrruy2267pV+/funWrVs22mijfOELX8h7771Xu2ZxLWpPP/10DjzwwPTp0yddu3bNNttsk8suu6zOmg9bua6++uqccsop6d+/f3r16pW99torzz33XP1+yEn23XffJMk111xTe+3NN9/Mn//85xxyyCGLfc5pp52WnXbaKX379k2vXr2y3Xbb5Y9//GNqampq1wwcODDPPPNM7r333tqf38CBA+uc/Yorrsj3vve9DBgwIF26dMm///3vRVrUZsyYkQ033DDDhw/P3Llza19/4sSJ6dGjR77yla/U+14BgI8fAYdSmD9/fsaNG5dNNtkk1dXV9XrOsccem5NPPjk77LBDLr/88nzve9/L7bffnl122SUzZsyos3batGn5n//5n3z+85/P5Zdfnp133jk//OEPM3bs2CTJLrvskltvvTVJstdee+XWW2+t/bq+Jk+enJEjR6Z169Y555xzcvXVV+fkk09OmzZt8v777y/xec8991x22WWXTJw4Maeffnouvvji9O/fP0cccUTOPvvsRdb/6Ec/yksvvZRf/vKXOeuss/L888/ngAMOyPz58+t1znbt2mXPPffMJZdcUnvtmmuuSYsWLbL33nsv9jkvvvhiDj300Fx00UW5+OKLs8cee+Tb3/52fvrTn9auueSSS9K3b99svPHGtT+//97jw7O//PLLOeOMM3LFFVekc+fOi+zVsWPHXHDBBXnkkUdy8sknJ0nee++9HHrooamurs6ZZ55Zr/sEAD6etKhRCjNmzMh7772XPn361Gv9s88+m4suuiiHH354fvazn9Ve33jjjbPTTjvl3HPPzfe///3a6zNnzsxVV12VwYMHJ0l22GGH3HPPPbn66qtz4IEHplOnTunUqVOSpEuXLsvVMvXYY49l9uzZ+dGPflRbuUiS/ffff6nPO+200/L+++/nz3/+c224Gz58eN58882cfvrpOfTQQ7P22mvXru/fv3/OO++82q9btmyZQw89NI888ki9z33IIYdkjz32yDPPPJMBAwbkkksuyV577bXE99/897CFBQsWZJtttklNTU1+85vf5IQTTkhRFBk0aFA+8YlPpG3btks8R9++ffOHP/xhmefbcsst8/3vfz8nn3xytt5669x4442ZPHlybrvttqy55pr1ukcA4ONJBYdV0oftTAcddFCd64MHD87666+fu+66q871rl271oabD2244YZ56aWXGu1MG2+8cVq3bp1jjjkml112WSZNmlSv5919993ZfvvtF6lcHXTQQXnvvffyj3/8o871ESNG1Pl6ww03TJIG3cs222yTfv365ZJLLslTTz2VRx55ZIntacnC1ru99torvXv3TocOHdKpU6f85Cc/ycyZMzN9+vR677vnnnvWe+1RRx2VXXbZJYcffnguv/zynH766bX3CgCUl4BDKXTs2DFt2rSp836WpZk5c2aShcHlo7p161b7/Q916NBhkXWtW7fOf/7zn+U47eL169cvf/rTn9KpU6ccf/zx2WSTTbLJJptkzJgxS33ezJkzl3gfH37/v330Xlq3bp0kDbqXoihy8MEH58orr8zvf//7fOpTn8rWW2+92LXjx4/PPvvskyQ5++yzc8stt+TOO+/Mcccd1+B9P7yn+p7xoIMOyuzZs9O1a9cccMAB9X4uAPDxJeBQCi1btsx2222Xxx57bJEhAYvz4S/506ZNW+R7r776ajp27NhoZ1tjjTWSJHPmzKlz/aPBI0m23nrrjB07Ni+++GJuu+22DBkyJCeeeGKdN/R/VIcOHZZ4H0ka9V7+20EHHZQZM2bkwgsvXOpY52uuuSarrbZaxo4dm7333jtbbLFFNt100+XasyiKeq999dVXc9xxx2XgwIGZOXNm/t//+3/LtScA8PEi4FAaxx57bGpqanL00Ucv9k35c+fOzV//+tckyXbbbZcktUMCPvTII4/kn//8Z7bffvtGO9eHn4Xz1FNP1bn+4VkWp2XLltlss81q3x80YcKEJa7dfvvtc/fdd+eVV16pc/2KK65ImzZtstlmmy3v0ZeqR48eOeqoo/LZz342Bx544BLXFUWRVq1apWXLlrXX/vOf/+SKK65YZG3r1q0ze/bsFT7b/Pnz8+UvfzlFUeTqq6/OSSedlPPOOy833HDDCr82ANC8GTJAaQwZMiRnnHFGvvWtb2WHHXbIYYcdlv79+2fevHl5/PHHc9FFF2XAgAEZMWJE1l133Rx66KE577zz0qJFi+y888558cUX8+Mf/zjV1dU54ogjGu1cO++8c9q3b58jjzwy3/3ud9OqVatcdtlli1SaLrzwwtx9990ZPnx4qqurM2fOnNopYjvssMMSX//b3/52brnlluyxxx454YQT0r59+1x11VW55ZZb8qMf/ajOgIHG9oMf/GCZa4YPH55f//rXOfzww3PooYdm5syZ+dWvfrXYUd4bbLBBrr322lx77bXp27dvVl999eV638ypp56a+++/P9ddd126du2a0aNH5957783o0aOz8cYbp2/fvg1+TQDg40HAoVRGjRqVT3/60zn33HNz1lln5bXXXstqq62WddZZJ/vtt1+dz0A544wz0q9fv1x88cU5//zz065du+y00045+eSTF/uem+XVrl27XHPNNTnxxBPz1a9+NWuvvXa+8IUv5DOf+UyOOuqo2nUDBw7MHXfckVNPPTWvvfZa1lxzzQwYMCCXX355hg0btsTXX3fddWvDzHHHHZfZs2dnvfXWy69//eulto5Vyvbbb59zzjknZ599dg444IB07949o0aNSqdOnTJ69Og6a0888cRMmzYtRx99dN5+++306tUrTzzxRIP2u/POO3PGGWfk+OOPr1OJGzNmTLbddtscdthhufnmm2vfewQAlEsxa9asmmUvAwAAaP68BwcAACgNAQcAACgNAQcAACgNAQcAACgNAQcAACgNAQcAACgNAQcAACgNAQcAACgNAQcAACgNAQcAACgNAQcAACgNAQcAACgNAQcAACgNAQcAACgNAQcAACgNAQcAACgNAQcAACgNAQcAACgNAQcAACgNAQcAACgNAQcAACgNAQcAACgNAQcAACgNAQcAACgNAQcAACgNAQcAACgNAQcAACgNAQcAACgNAQcAACgNAQcAACgNAQcAACgNAQcAACgNAQcAACgNAQcAACgNAQcAACgNAQcAACgNAQcAACgNAQcAACgNAQcAACgNAQcAACgNAQcAACgNAQcAACgNAQcAACgNAQcAACgNAQcAACgNAQcAACgNAQcAACgNAQcAACgNAQcAACgNAQcAACgNAQcAACgNAQcAACgNAQcAACgNAQcAACgNAQcAACgNAQcAACgNAQcAACgNAQcAACgNAQcAACgNAQcAACgNAQcAACgNAQcAACgNAQcAACgNAQcAACgNAQcAACgNAQcAACgNAQcAACgNAQcAACgNAQcAACiNVk19gIba5qjL88bbc5r6GABNZvzvv97URwBoQjVZI7Ob+hDN3rajfpmZb75bsf06rL1mxv3hqIrttzQfu4DzxttzMuMt/6MGVmFF0dQnAGg6NU19gI+HmW++mxlvvtfUx2gSWtQAAIDS+NhVcAAAgGUoispW/JtRd4EKDgAAUBoqOAAAUDZFi4WPSu7XTDSfkwAAAKwgAQcAACgNLWoAAFA2hgwAAAB8/KngAABA2RgyAAAA8PEn4AAAAKWhRQ0AAMrGkAEAAICPPxUcAAAoG0MGAAAAPv4EHAAAoDS0qAEAQNkUqfCQgcpttSwqOAAAQGmo4AAAQNkYMgAAAPDxJ+AAAACloUUNAABKp6jskIFmNGVABQcAACgNFRwAACgbQwYAAAA+/gQcAACgNLSoAQBA2RQVHjJQ0YEGS6eCAwAAlIYKDgAAlI0hAwAAAB9/Ag4AAFAaWtQAAKBstKgBAAB8/Ak4AABAaWhRAwCAsmlRLHxUcr9mQgUHAAAoDRUcAAAom6Ko8JABFRwAAIBGJ+AAAACloUUNAADKpigq2zamRQ0AAKDxqeAAAEDZFC0qPGSg+dRNms9JAAAAVpCAAwAAlIYWNQAAKBtDBgAAAD7+VHAAAKBsDBkAAAD4+BNwAACA0tCiBgAApVPhIQMxZAAAAKDRqeAAAEDZGDIAAADw8SfgAAAApaFFDQAAyqao8JCBig40WDoVHAAAoDRUcAAAoGyKosJDBlRwAAAAGp2AAwAAlIYWNQAAKBtDBgAAAD7+VHAAAKBsihYVHjLQfOomzeckAAAAK0jAAQAASkOLGgAAlI0WNQAAgI8/FRwAACgbY6IBAAA+/gQcAACgNLSoAQBA2RgyAAAA8PEn4AAAAKWhRQ0AAMrGFDUAAICPPxUcAAAom6Ko8JABFRwAAIBGJ+AAAACloUUNAADKxpABAACAjz8VHAAAKJmiKFJUsKpSyb2WRQUHAAAoDQEHAAAoDS1qAABQOpVtUUu0qAEAADQ6FRwAACibIpUtqjSfAo4KDgAAUB4CDgAAUBpa1AAAoGR8Dg4AAEAJqOAAAEDJqOAAAACUgIADAACUhhY1AAAoGS1qAAAAJaCCAwAAJaOCAwAAUAICDgAAUBpa1AAAoGyKDx6V3K+ZUMEBAABKQwUHAABKxpABAACAEhBwAACA0tCiBgAAJaNFDQAAoARUcAAAoGRUcAAAAEpAwAEAAEpDixoAAJSMFjUAAIASUMEBAICyKT54VHK/ZkIFBwAAKA0BBwAAKA0tagAAUDJFKjxkoBn1qKngAAAAK93zzz+f4cOHZ/DgwRk2bFgmTpy4yJqampp8//vfz5Zbbpmtt946u+++e/797383aB8BBwAAWOmOOeaYjBo1KuPHj8/RRx+d0aNHL7Lmpptuyn333Zdx48blvvvuy/bbb58f/ehHDdpHwAEAgJIpiv/7LJzKPJZ+nunTp2fChAkZOXJkkmTPPffM5MmTM3ny5EXWzpkzJ7Nnz05NTU3efvvt9OjRo0H37j04AADASjVlypR07949rVotjB9FUaS6ujovv/xy+vTpU7tuxIgRueeee7L++utnrbXWSvfu3XPjjTc2aC8VHAAAKJnKVm/qN9Dgo2tqamoWWTNhwoQ899xzefrppzNx4sRsv/32Of744xt07wIOAACwUvXs2TNTp07NvHnzkiwMN1OmTEl1dXWddZdddlm23XbbVFVVpUWLFjnwwAMzbty4Bu0l4AAAACtV586dM3DgwIwdOzZJcsMNN6R379512tOSpE+fPrnrrrsyd+7cJMnNN9+cDTbYoEF7eQ8OAACUTfHBo5L7LcNZZ52VI444ImeccUbatm2bMWPGJElGjx6dESNGZNddd83//M//5Nlnn83WW2+d1VZbLd26dcuZZ57ZoKMIOAAAwEq37rrr5tZbb13k+q9+9avav6+++ur55S9/uUL7CDgAAFA29Xzjf2Pu11x4Dw4AAFAaAg4AAFAaWtQAAKBk6vvZNI25X3OhggMAAJSGCg4AAJSMCg4AAEAJCDgAAEBpaFEDAICS0aIGAABQAio4AABQNsUHj0ru10yo4AAAAKUh4AAAAKWhRQ0AAErGkAEAAIASUMEBAICSUcEBAAAoAQEHAAAoDS1qAABQMlrUAAAASkAFBwAAyqj5FFUqSgUHAAAoDQEHAAAoDS1qAABQMoYMAAAAlIAKDgAAlIwKDgAAQAkIOAAAQGloUQMAgJLRogYAAFACKjjwgS/vvmlG7zckXTuslYmTX893f3N77n/q5SWuP3yPTXP4HoPTu2u7vDz9rfzi8vsz9vanKnhigIa7+Lp78tsr7sxrM9/Ken275aQjP5chg9ZZ4voHHvtX/vfX1+fZSa+ma8d2+eqBw3LIXkNrvz933vyce8ltueaWf+TV19/MJ3t1yXe+unt22GJAndd5dfqsnPbbv+TvDz6T2XPmpl+vzvnpCQdk4Pq9Vtq9AqsmAQeS7L1d//zkqzvluF//LQ8+NSWH7rpJrvzf/bPVV87Py9PfXmT9Ybttku8fun2OOfvmPPLsKxm8fo+cdfQuefOd2bn5weeb4A4Alu3PdzyaH53zp5zyzf2y2Ub9cumf78uh3z4vt/7hO+nZtf0i6196ZUa+9O3f5YDdt8xZ3zskDz/5Qr5/5tXpWLVWRmw/KEny8/Nvyp9uHZ/Tjv981undJXc99M989f/9Ptf8+qhstF51kuTNt9/Lvkf+Mlttsm4u+ulX0rGqbV6c+nrarfWJit4/rEqKorJtY82oQ61pW9Sef/75DB8+PIMHD86wYcMyceLEpjwOq7Aj9tk8l9zyeC6++fE8+9KMfPe3t2fK9Ldz2O6bLnb9yJ02yh/++liuu3tiJr/6Zq6965lccssTOfrzW1b45AD1d/6Vf8/nd90iB+y+ZT7Vt2tOHr13uneuyiXX37vY9Zdcf196dKnKyaP3zqf6ds0Bu2+Z/XcdkvOuuLN2zXV/ezjfOOQz2XHLDdK7R6d84XNDs92Q9XP+lX+vXTPmstvTo3NVfn7igdlkQJ/06t4hQwevlz49O63sWwZWQU0acI455piMGjUq48ePz9FHH53Ro0c35XFYRa3WqkU2Wbdb7njkhTrX73zkhQwZ0HOxz2m9WsvMfn9+nWuz35+bT6/XPa1aemsb0Py8P3dennz25Wy7+fp1rm+7+foZ/+SkxT7n0acmLbJ+u83754l/vpS58+bXvu7qres2hKzRerX844l/1359271PZWD/XjnipIsyeK/vZ9cv/zyX//n+RrgrYImKJng0E032m9j06dMzYcKEjBw5Mkmy5557ZvLkyZk8eXJTHYlVVMd2bdKqZYtMf+O9Otenv/FuunRYc7HPuWP8C/nCZzfOoE91TZJssm63HDx847RerWU6rq3lAmh+3njz3cyfvyCdO7Stc71z+7Z5feZbi33O9Jlvp3P7j6zv0Dbz5i/IG2++k2Rh4Dn/yr/nhZenZ8GCBRn3j3/m1nufzPQZ//eaL74yI5dcf1/6VnfOH3721Ry819b5wS+vyzU3/6OR7xKgCd+DM2XKlHTv3j2tWi08QlEUqa6uzssvv5w+ffo01bFYhdWkps7XRVGkpmbxa3922X3p0n7N3HrWF1IURV57491cfuvCFrX585fwJIBmoe4/s9YkS2+e/8j3amr/w7jw+slH7Z3v/GxsdvrCqSmKIn16dMz+I4bkqr8+9H/PWVCTgev3yglf2S1JstF61XnuhVdzyfX3Zt/Pbr6iNwRQR5MOGfjoG59qlvTbJKxEM956L/PmL0iX9nWrNZ2q2mT6G+8u9jmz35+X0Wf+Nd/85S3p0n7NvDrznRw6YlDeendOZrz13mKfA9CU2q+9Zlq2bJHpH6nWvP7G2+n0kSrNhzp3aLuY9e+kVcsWab/2wv9mdqxaK7/78Zcze87czHrr3XTttHZO++1f0qt7h9rndOnYLuv27Vrnddbp0zV/vfvxxrg1YDF8Dk4T6NmzZ6ZOnZp58+YlWRhupkyZkurq6qY6EquoufMW5LHnXs2Om/atc32HTfvmoWemLPW58+YvyNTX386CBTXZZ/sB+dtDzy+x6gPQlFqv1iobrVedex5+ts71ex5+NoM36rvY52y6Yd9F1o/7xz8zcP1eWa1VyzrX11h9tXTrXJV58xfk5rsfz85DB9Z+b/BG/fLvF1+rs/6Fl19b7OQ2gBXVZAGnc+fOGThwYMaOHZskueGGG9K7d2/taTSJc6/9R77w2UE5ePjArNerY378lWGp7tIuv7/xsSTJSV/aLmOO2612/To92+fzwzbIJ3u0z6fX654LvrNnBvTtnB9ddFcT3QHAsh3++R0y9sYHcuWND+Zfk6blR+dcl6mvvZGD99w6SXL6eX/JsT++tHb9IXttnSnT3sgp5/wp/5o0LVfe+GCuvOnBfOWAHWvXPPr05Nx89+N5cerreWjC8xl1/G+zYMGCfPXAYbVrvrz/9nn06cn59cW3ZtLL03P9reNz+Z8fyBf33qZyNw+rmg8qOJV6NKc50U3aonbWWWfliCOOyBlnnJG2bdtmzJgxTXkcVmHX3T0xHdp9IiccPDRd26+ZZya/npHfvyovvbawNaNrh7VS3aVd7fqWLVrkG/sMyaeqO2Te/AUZN2Fydjn2krw0bfFv1AVoDvYYtmlmvfluzv7jLZk+462s1697fn/6V1LdbWE72Wsz3sqU196oXd+re8f8/vT/ySnn/CkX/+medOm4dk4+au/az8BJkjnvz83Pz78pL74yI2t+YvXsuMWAnPm9g7N22/8buDJoQO/89n8Py0/PuzFn//Fv6dWtQ0468nP53M6DK3fzwCqjmDVr1seqoWbDL12UGW/NbupjADSZSdce29RHAGg6NTVZI/9p6lM0e8N+/lBmvTevYvtVtWmVO44bUrH9lqZJKzgAAEDjq3TXWDPqUGvaD/oEAABoTCo4AABQMsZEAwAAlICAAwAAlIYWNQAAKBlDBgAAAEpABQcAAErGkAEAAIASEHAAAIDS0KIGAAAlU6TCQwYqt9UyqeAAAACloYIDAAAl06JFkRYtKldXqeRey6KCAwAAlIaAAwAAlIYWNQAAKJmiqPCQgebToaaCAwAAlIcKDgAAlExRFCkqWFap5F7LooIDAACUhoADAACUhhY1AAAoGUMGAAAASkAFBwAASsaQAQAAgBIQcAAAgNLQogYAAGVT4Ra15jRlQAUHAAAoDRUcAAAoGWOiAQAASkDAAQAASkOLGgAAlIzPwQEAACgBAQcAACgNLWoAAFAypqgBAACUgAoOAACUjCEDAAAAJSDgAAAApaFFDQAASqZIhYcMVG6rZVLBAQAASkMFBwAASsaQAQAAgBIQcAAAgNLQogYAACVTFBUeMtB8OtRUcAAAgPJQwQEAgJJZWMGp5JCBim21TCo4AABAaQg4AABAaWhRAwCAkjFkAAAAoARUcAAAoGSKoqjwkIHmU8JRwQEAAEpDwAEAAEpDixoAAJSMIQMAAAAloIIDAABlU+EhA82phKOCAwAAlIaAAwAAlIYWNQAAKBlDBgAAAEpABQcAAEqmqPCQgYoONFgGFRwAAKA0BBwAAKA0tKgBAEDJaFEDAAAoARUcAAAoGWOiAQAASkDAAQAASkOLGgAAlEyRCg8ZSPPpUVPBAQAASkPAAQAASkOLGgAAlIwpagAAACWgggMAACVTFBUeMtCMSjgqOAAAQGkIOAAAQGloUQMAgJIxZAAAAKAEVHAAAKBkWhRFWlSwrFLJvZZFBQcAACgNAQcAACgNLWoAAFAyhgwAAACUgAoOAACUTVGkWEVLOCo4AABAaQg4AADASvf8889n+PDhGTx4cIYNG5aJEycudt1TTz2V3XbbLUOGDMlmm22WG264oUH7aFEDAICSaVEsfFRyv2U55phjMmrUqBx88MG5/vrrM3r06Nx666111rz33ns5+OCDM2bMmGy11VaZN29eZs2a1bCzNGg1AABAA02fPj0TJkzIyJEjkyR77rlnJk+enMmTJ9dZd/XVV2fIkCHZaqutkiStWrVKp06dGrSXgAMAACVTfDBkoJKPpZkyZUq6d++eVq1a1Z6vuro6L7/8cp11EydOzOqrr56RI0dmm222yVe/+tW8/vrrDbp3AQcAAFjpPhqCampqFlkzd+7c3HHHHTnzzDMzbty4VFdX57jjjmvQPgIOAACwUvXs2TNTp07NvHnzkiwMN1OmTEl1dXWddb17984222yTHj16pCiK7L///hk/fnyD9hJwAACgZIqi8o+l6dy5cwYOHJixY8cmSW644Yb07t07ffr0qbPuc5/7XB599NG89dZbSZLbbrstG220UYPu3RQ1AABgpTvrrLNyxBFH5Iwzzkjbtm0zZsyYJMno0aMzYsSI7LrrrunVq1e++c1vZvjw4WnZsmV69OiRs846q0H7FLNmzVq0+a0Z2/BLF2XGW7Ob+hgATWbStcc29REAmk5NTdbIf5r6FM3e1657Pm/PmV+x/dqu3jK/2Xudiu23NFrUAACA0hBwAACA0vAeHAAAKJkWxcJHJfdrLlRwAACA0lDBAQCAkimKYpEP1lzZ+zUXKjgAAEBpCDgAAEBpaFEDAICSKZJUsmus+TSoqeAAAAAlooIDAAAls3BMdOXqKsZEAwAArAQCDgAAUBpa1AAAoGSKosJDBrSoAQAAND4VHAAAKJuiSLGKlnBUcAAAgNIQcAAAgNLQogYAACVjyAAAAEAJCDgAAEBpaFEDAICSaVEUaVHBvrFK7rUsKjgAAEBpqOAAAEDJFB88Krlfc6GCAwAAlIaAAwAAlIYWNQAAKJmiKFJU8I3/ldxrWeoVcO699956v+DQoUOX+zAAAAArol4BZ/fdd69XKiuKIjNmzFjhQwEAAMuvRbHwUcn9mot6t6jV1NSszHMAAACssHoFnAkTJqzscwAAAKywegWc3r17L/F78+bNS6tWZhUAAEBzsSoPGViuMdH33HNPdt1113Tt2jW77bZb7rrrrnzjG9/Igw8+2NjnAwAAqLcGl17GjRuXffbZJ/PmzUuy8L051dXVueyyy5IkW2yxReOeEAAAaJCiWPio5H7NRYMrOD/5yU8yf/787L777rXX1llnnXTp0kUFBwAAaFINDjiPPfZY+vTpk4svvrjO9a5du+aVV15ptIMBAAA0VINb1Fq1arXIyOgFCxbklVdeScuWLRvtYAAAwPIxZKABBg4cmBdffDFHHXVUkuT111/Pl7/85bz++usZNGhQox8QAACgvhoccL75zW8mSS655JIURZFJkybl+uuvT1EUtaEHAABoOkWRtKjgoxkVcBoecHbeeeecf/75qa6uTk1NTe0UtfPOOy8777zzyjgjAABAvSzXJ3Tus88+2WeffTJjxozU1NSkU6dOjX0uAACABluugDN79uxcccUVeeqpp5IkG220UUaOHJk11lijUQ8HAAA03Ko8ZKDBAefpp5/O5z//+UydOrXO9Z/+9Ke58sors+GGGzba4QAAABpiuYYMTJkyJTU1NVl77bWz9tprp6amJlOnTs23vvWtlXFGAACgAYomeDQXDQ44EyZMSOvWrXPNNdfkhRdeyAsvvJBrr702rVu3zmOPPbYSjggAAFA/DW5R69WrV1ZbbbUMGzas9tqOO+6YT37yk1mwYEGjHg4AAKAhGlzB+cEPfpBJkybl73//e+21O+64I5MnT87JJ5/cmGcDAACWQ4sUaVFU8NGMmtTqVcEZNGhQna8XLFiQffbZJ1VVVSmKIm+88UZat26d7373u9l1111XykEBAACWpV4B58UXX1zs9TfeeKP273PmzFniOgAAoHKKYuGjkvs1F/UKON/+9rdX9jkAAABWWL0Czne+852VfQ4AAIAV1uApah+aNGlSXn311cyfP7/O9aFDh67woQAAgOVXFEWKCvaNVXKvZWlwwJk2bVoOPvjgPPLII4t8ryiKzJgxo1EOBgAA0FANDjg/+MEPMn78+JVxFgAAoBGsykMGGvw5OH//+9/TokWLnH322UmS/v375+STT0779u3z+9//vtEPCAAAUF8NDjivv/561l133Xzxi19Mkqy55po55phj0rlz51xzzTWNfkAAAID6anCLWps2bdKyZcskC8PNpEmT8tprr+X111/PlClTGv2AAABAw7QoirSoYN9YJfdalgZXcHr06FEbZNZZZ53MnDkz/fv3zxtvvJG111670Q8IAABQXw0OOJ/5zGfSo0ePPP300/n617+eJKmpqUlNTU2+9rWvNfoBAQAA6qvBLWqnnHJKTjnllCTJBhtskD59+mT8+PHZaKONssMOOzT2+QAAgAZalaeoLfcHfX5oq622ylZbbdUYZwEAAFgh9Qo4e+yxR71erCiK3HDDDSt0IAAAYMUURZGigmWVSu61LPUKOPfcc0+KokhNTc1S1zWnGwMAAFY99Qo4BxxwQLMJL8+NHZ2iaPBsBIDSaL/5kU19BIAm03HtNnnq+u819TFoxuoVcMaMGbOyzwEAADSSIssxLnkF92sulEIAAIDSWOEpagAAQPOyKg8ZUMEBAABKQ8ABAABKQ4saAACUTIti4aOS+zUXyxVw5syZk6uuuioPP/xwunTpki984Qt58cUXs8EGG6R9+/aNfUYAAIB6aXDAmTlzZnbfffdMnDgxSbLZZptliy22yP7775/jjz8+J554YqMfEgAAqL9VuYLT4PfgnHTSSXnmmWeyxhprpKamJkmyww47pE2bNrntttsa/YAAAAD11eCAc8stt6Rdu3Z59NFHa6+1bNkyvXr1yqRJkxrzbAAAAA3S4Ba1t956K+utt166du1a5/r8+fPzzjvvNNrBAACA5VMUlf1smmb0MTgNr+BUV1dn4sSJuf/++2uv/fWvf81zzz2X3r17N+rhAAAAGqLBAWfffffNvHnzsttuu6Uoijz88MM5+OCDUxRF9t1335VxRgAAoAE+HDJQyUdz0eCAc9xxx2XnnXdOTU1NncewYcNy7LHHrowzAgAA1EuD34PTunXrXHnllbn33nvzyCOPpKamJp/+9KezzTbbrIzzAQAA1NtyfdBnkgwdOjRDhw5tzLMAAACNoEhl3/jfjDrUGh5w9thjjyV+ryiK3HDDDSt0IAAAgOXV4IBzzz33pCiK2g/5TFL7dSVH0QEAAItXFEVF3/jfnHJAgwPOAQccUOcG3nrrrdxzzz155513ss8++zTq4QAAABqiwQFnzJgxi1ybMWNGhg4dmp49ezbKoQAAAJZHg8dEL07Hjh3Tr1+/XHbZZY3xcgAAwApo0QSP5qLBFZzTTz+9ztfz58/P888/nwceeCBVVVWNdS4AAIAGa3DAOe2005b4JqJddtllhQ8EAACsmKKo8Jjo5jNjoOEBp7q6uk7AKYoinTt3zvbbb59vfvObjXo4AACAhmhwwHniiSdWxjkAAABWWIMCzty5c7P55punqqoqd955Z7Oadw0AACzUosKfg9OiGeWCBg08WG211fLOO+9k7ty5wg0AANDsNHii20EHHZR//etfefLJJ1fGeQAAgBX04ZCBSj6aiwa/B2fatGlJkp122inbbrttunTpUlvNKYoi55xzTuOeEAAAoJ4aHHCuvPLKFEWRmpqa3H777bXhpqamRsABAACaVL0Dzumnn54ePXpk6NChK/M8AADACmpRpMJDBiq317LUO+Ccdtpp2XzzzfO3v/1tZZ4HAABguTW4RQ0AAGjeVuUx0Q0KOHPmzMlLL7201DW9evVaoQMBAAAsrwYFnCeeeCKDBg1a4veLosiMGTNW+FAAAADLo8EtajU1NSvjHAAAQCOp9GfTNKMOtYYFnB49euSQQw5ZWWcBAABYIQ0OON/5zndW1lkAAABWiClqAABQMkWFPwenObWotajvwurq6nTt2nVlngUAAGCF1LuC88QTT6zMcwAAAI2kSJFKFlUqu9vS1buCAwAA0NwJOAAAQGkYMgAAACXTosJDBiq517Ko4AAAAKWhggMAACWjggMAAFACAg4AAFAaWtQAAKBkiqJIUcG2saKSmy2DCg4AAFAaKjgAAFAyLVLhIQOV22qZmtNZAAAAVoiAAwAAlIYWNQAAKJmiSIWHDFRur2VRwQEAAEpDBQcAAEqmKIqKDhkwJhoAAGAlEHAAAIDS0KIGAAAl06Ko8OfgNJ8ONRUcAACgPFRwAACgZIyJBgAAKAEBBwAAKA0tagAAUDItUlS0ktEizadHTQUHAABY6Z5//vkMHz48gwcPzrBhwzJx4sQlrp09e3a22GKL7LDDDg3eR8ABAICS+XDIQCUfy3LMMcdk1KhRGT9+fI4++uiMHj16iWtPOeWUbL755st17wIOAACwUk2fPj0TJkzIyJEjkyR77rlnJk+enMmTJy+y9r777svzzz9fu7ahBBwAAGClmjJlSrp3755WrRaOACiKItXV1Xn55ZfrrHv33Xdz4okn5owzzljuvQQcAAAomRZF5R/LUnykj62mpmaRNSeddFIOP/zw9OjRY7nv3RQ1AABgperZs2emTp2aefPmpVWrVqmpqcmUKVNSXV1dZ93999+fv/3tb/npT3+aOXPmZNasWdlyyy3zwAMP1HsvFRwAACiZFkVR8cfSdO7cOQMHDszYsWOTJDfccEN69+6dPn361Fl333335YknnsgTTzyRCy64IBtssEGDwk0i4AAAABVw1lln5aKLLsrgwYNz5pln5le/+lWSZPTo0bnpppsabZ9i1qxZiza/NWOt27RNUchlwKqr/eZHNvURAJpMx7Xb5Knrv9fUx2j2/vTsm5kzv3K/5q/essjn1lu7YvstjffgAABAydT3s2kac7/mQikEAAAoDQEHAAAoDS1qAABQMkVR1OuzaRpzv+ZCBQcAACgNFRwAACgZQwYAAABKQMABAABKQ4saAACUTItUtpLRnKomzeksAAAAK0QFBwAASqYoigoPGWg+UwZUcAAAgNIQcAAAgNLQogYAACVTfPCo5H7NhQoOAABQGio4AABQMi1SpEUFyyotmlENRwUHAAAoDQEHAAAoDS1qAABQMoYMAAAAlIAKDgAAlExRLHxUcr/mQgUHAAAoDQEHAAAoDS1qAABQMgtb1CrXN6ZFDQAAYCVQwQEAgJIpUtlKRjMq4KjgAAAA5SHgAAAApaFFDQAASqYoCp+DAwAA8HGnggMAACVTpLJv/G9GBRwVHAAAoDwEHAAAoDS0qAEAQMkYMgAAAFACKjgAAFAyLVLZSkZzqpo0p7MAAACsEAEHAAAoDS1qAABQMoYMAAAAlIAKDgAAlEzxwaOS+zUXKjgAAEBpCDgAAEBpaFEDAICSKYrKvvHfkAEAAICVQMABAABKQ4saAACUTIsUFa1kNKeqSXM6CwAAwApRwQEAgJIxZAAAAKAEBBwAAKA0tKgBAEDpFGlGXWMVpYIDAACUhgoOAACUjCEDAAAAJSDgAAAApaFFDQAASqZFiopWMppT1aQ5nQUAAGCFqOAAAEDJGDIAAABQAgIOAABQGlrUAACgZLSoAQAAlIAKDgAAlEyRIpUsqjSjAo4KDgAAUB4CDgAAUBpa1AAAoGRaFAsfldyvuVDBAQAASkMFBwAASsaQAQAAgBIQcAAAgNLQogYAACVTJCkq2DemRQ0AAGAlUMEBAICSMWQAAACgBAQcAACgNLSoAQBAyRRFZSsZlRxosCwqOAAAQGmo4AAAQMkYMgAAAFACAg4AAFAaWtQAAKBkiqKybWOGDAAAAKwEKjisEs6/6u786pLbM+31N9P/k93zk2P3zdabfmqJ6+8d/1y+d9a1mfjvV9Kt09o56oufyWH7blv7/T9cd2+uuOmhPPP81CTJJv175/vf2CODN+xbu+aM39+Sv9w5Ic9NnpY1Vl8tQzb+ZH5w5F5Zt2/XlXafAMvjy/ttm9GH7JSundbOxH+/ku+ecU3uf+z5Ja4/fP/tcvj+26V39w55edob+cWFt2TsTQ/VWfO1A3fIYftum+qu7TPzzXdz/e2P5ke/viFz3p+3sm8HWMWp4FB61/5tfL57xjX51pd2yV2XfCdbbbJOPn/0uXnp1ZmLXT95yuv5/DFjstUm6+SuS76TY7+0S77z86tzwx2P1q65Z/xz2Xf44Px5zNH524XfSnW39tnnyF9n6muzatfc98i/cvj+2+VvFx6Xa885MvPmz88+o8/Ju/+Zs7JvGaDe9t750/nJsfvmF7+/Jdsfclruf+z5XHn2Eanu2n6x6w/bd5t8/4g9cvrvbspWB/w4p/32pvzshM/ns9tuVLtm/89ulpO/sVd++ru/ZovP/29Gn3Jp9t55cE76xp6Vui1Y5RVN8GgumjTgnHDCCRk4cGCqqqry9NNPN+VRKLFzL7sjh+y1Vb74ua2zfr9uOfVb+6Vn1/a58Opxi11/4bX3pLpb+5z6rf2yfr9u+eLnts7Be26Zcy65vXbN7/730By+/3YZuH511uvbLWd/76DU1NTk7n/8s3bN1b/6Rg7aY8sMWKd7Bq5XnV+fdEhefvWNPPbMSyv9ngHq64iDhuWS6+/Pxdffn2cnTct3z7gmU6a9kcP223ax60fuOiR/uO7eXHfrI5k8ZUauvXV8Lrnh/hz9xZ1r12w+sF8efPzfufqWh/PSKzNz54MTc83fHs6mA3pX6raAVViTBpy99torN998c3r16tWUx6DE3p87L49NfCnDthhQ5/qOWwzIQ4+/sNjn/OOJF7LjR9bvtOUGefTpFzN33vzFPue92e9n7rz5qWrXZolneeud2UmS9ktZA1BJq7VqmU3698odDz5T5/qdDz6TIRv3W+xzWq/WKrPfn1vn2uw5c/PpDfukVcuFv1Y88Ni/s0n/Xvn0Bn2SJH16dszOW2+Yv9371Eq4C2BxWhRFxR/NRZO+B2fo0KFNuT2rgBmz3sn8+QvSuUPbOtc7d2yb12a8tdjnvDbjrXTu+JH1Hdpm3vwFmTHrnXTrtPYiz/nhOdene+e1s8OQ/ot9zZqamnzvzGuy5SbrZINP9VjOuwFoXB2r1kqrVi0zfebbda5Pn/F2unRst9jn3PHAM/nCXlvnxr8/ngkTX8omA3rn4D22TOvVWqVj1VqZNuOtXHvr+HRsv1b+ev43UxRFVmvVMhdcfXfO+sOtlbgtYBVnyACrhI/+o0JNTU2KpfxLw0e/U5OaD64v+pyz/3hrrvnb+Pz5N0dnjdVXW+zrHf/TK/PUv6bmr7/7ZoPODVAJNTV1vy6KIjUfvfiBn11wc7p0bJdbf39ciiSvzXw7l//lwRw9aufMX7AgSTL00+vmW4ftkuNOH5vxT05Ov16dctq39surr7+Vn19w80q+G2BVJ+BQah2r1krLli3y2oy6/zr5+sx3FqnqfKhLx3aLXd+qZYt0qFqzzvVfXXxbzvj93/KnXx+ZjdbtudjXO+FnV+avdz+Rm847Jj2X8KZdgKYwY9Y7mTdvfrp8pGrdqcNai1R1PjR7ztyMPuXSfPMnl6dLx3Z59fU3c+jeQ/PWO//JjFnvJkm+97XdcuVND+Xi6+9Pkjz9/NSs+YnVc+Z3D8wvLrxlieEJaDyVfuN/82lQM0WNkmu9Wqts0r9X7nxwYp3rf39o4hL7yzcf2C9/f6ju+jsefCabbtA7q7VqWXvtlxfflp9dcHOu/uUR2fSDPvP/VlNTk+N/emX+cueE3DDmqPTp2akR7gig8cydNz+PTXwpO25Rt712hyH9l/g+xQ/Nm78gU1+blQULarLP8MH52z1P1QaXT6zROgsW1A0x8+cvWPgLV3P6LQgoJRUcSu+Ig4blayf/MZtu0DubD+yXP1x3b15+dWa+9MHn2vzwnOvzyvQ385sffjFJctg+2+T8K+/O9868Jl/83ND844kXcsn19+f8Hx9a+5pn//HW/OQ3N+Z3/zsqvbt3zLTXF76fZ802q2etNqsnSY47/cpcfcvDueznX8labdaoXdNurTXyiTVaV/AnALBk5152R37zwy/m0adfzD+eeCGj9h6a6m4d8vtrFk6aPOkbe6Z757Xz9R9cnCRZp3eXDN6wTx5+clKq2rbJNw4elgGf7FH7/SS5edyTOeKgHfP4P1/Ow09NyierO+e7X9s9fx33xCLBB1hJKv2PCc3oHy+aNOAcd9xxuemmmzJt2rR87nOfy5prrplHH3102U+EBthn+ODMfPPd/PT8v2ba629lwDrdM/asI9K7e4ckybTX38rL//WZOH16dsqVZ3093z3zmpx/1bh067x2Tjtuv+w5bNPaNRdcPS7vz52XUd++oM5e3/6fEfnOV3ZLklz4wS8Hu3/t7Dprfn3SITlojy1Xyr0CNNR1tz6SDmuvmRMOH5Gundrlmedfychjzs1Lr76RJOnaqV2qu3WoXd+yRZFvHDwsn+rTNfPmzc+4h5/NLof/Ii+98n//Hf35hTcvHK7y9d3TvfPamTHrndw87smccu6fK35/wKqnmDVr1sfqn1Jat2mbotBZB6y62m9+ZFMfAaDJdFy7TZ66/ntNfYxm79kZNZlfwd/yWxbJeh2bRxlHixoAAJRMsdjZrytzv+ZDKQQAACgNFRwAACiZoqjwmOhmVMJRwQEAAEpDwAEAAEpDixoAAJRMkQq3qFVwr2VRwQEAAEpDBQcAAMqm0iWVZlTCUcEBAABKQ8ABAABKQ4saAACUzMIhA5XrGytSU7G9lkUFBwAAKA0VHAAAKJuiwu/7N2QAAACg8Qk4AABAaWhRAwCAkilS2a6xZtShpoIDAACUhwoOAACUTaVLKs2ohKOCAwAAlIaAAwAAlIYWNQAAKJnigz+V2y9Jaiq239Ko4AAAAKWhggMAACVTFBUeE23IAAAAQOMTcAAAgNLQogYAACVTpMItahXca1lUcAAAgNJQwQEAgLKpdEmlGZVwVHAAAIDSEHAAAIDS0KIGAAAlU3zwp3L7JUlNxfZbGhUcAACgNAQcAABgpXv++eczfPjwDB48OMOGDcvEiRMXWXPXXXdlp512yhZbbJGtttoqp5xySmpqGlYZEnAAAKBkiqLyj2U55phjMmrUqIwfPz5HH310Ro8evciaqqqqXHDBBXnwwQdz55135t57783VV1/doHsXcAAAgJVq+vTpmTBhQkaOHJkk2XPPPTN58uRMnjy5zrpBgwalb9++SZI11lgjAwcOzKRJkxq0l4ADAAAlUzTBY2mmTJmS7t27p1WrhTPOiqJIdXV1Xn755SU+Z9q0abn++uuz8847N+jeBRwAAGClKz7Sx7a099a89dZbOeCAA3LUUUdlk002adA+Ag4AALBS9ezZM1OnTs28efOSLAw3U6ZMSXV19SJr33777ey3334ZMWJEjjzyyAbvJeAAAEAZNZf+tCSdO3fOwIEDM3bs2CTJDTfckN69e6dPnz511r3zzjvZb7/9MmzYsJxwwgnLddsCDgAAsNKdddZZueiiizJ48OCceeaZ+dWvfpUkGT16dG666aYkyW9+85uMHz8+f/nLX7LNNttkm222yc9//vMG7VPMmjWreXzkaD21btM2RSGXAauu9ps3vFwPUBYd126Tp67/XlMfo9mb+k7LLKhPaaWRtEhNeqw1v2L7LY2kAAAAlIaAAwAAlEarpj4AAADQuIqiXu/9b7z9KrjXsqjgAAAApaGCAwAAJVPP6c2Nul9zoYIDAACUhoADAACUhhY1AAAom+bUM1ZhKjgAAEBpqOAAAEDJFB/8qdx+zYcKDgAAUBoCDgAAUBpa1AAAoGyK5tU2VkkqOAAAQGmo4AAAQMlUunrTnKpFKjgAAEBpCDgAAEBpaFEDAICyaU49YxWmggMAAJSGCg4AAJRMkSKVLOM0p4KRCg4AAFAaAg4AAFAaWtQAAKBkigr3jGlRAwAAWAlUcAAAoGQqXVFRwQEAAFgJBBwAAKA0tKgBAEDZNKeesQpTwQEAAEpDwAEAAEpDixoAAJRMkSKV7FNrTh1xKjgAAEBpqOAAAEDJFBUuqajgAAAArAQCDgAAUBpa1AAAoGQq3TKmRQ0AAGAlUMEBAICyaU4llQpTwQEAAEpDwAEAAEpDixoAAJRMkSKV7FNrTh1xKjgAAEBpqOAAAEDJFBUuqajgAAAArAQCDgAAUBpa1AAAoGQq3TKmRQ0AAGAlUMEBAICSMWQAAACgBAQcAACgNLSoAQBA6RRpXo1jlaOCAwAAlIYKDgAAlE2xqtZvVHAAAIASEXAAAIDS0KIGAAAls6q2pyUqOAAAQImo4AAAQMkUq3AJRwUHAAAoDQEHAAAoDS1qAABQMkWKrKqjBlRwAACA0lDBAQCAMqpkAaemgnstgwoOAABQGgIOAABQGlrUAACgZFbN8QILqeAAAACloYIDAAAlUzTFlOhmMmhABQcAACgNAQcAACgNLWoAAFAyRZqiR615UMEBAABKQ8ABAABKQ4saAACUjSlqAAAAH38qOAAAUDKr5niBhVRwAACA0hBwAACA0tCiBgAAJVOsuh+Do4IDAACUhwoOAACUTNEkc6KbBxUcAACgNAQcAACgNLSoAQBAyRgyAAAAUAICDgAAUBoCDgAAUBoCDgAAUBqGDAAAQMkYMgAAAFACKjgAAFAyxX/938qoqeBeS6eCAwAAlIaAAwAAlIYWNQAAKJvig0EDFdJ8GtRUcAAAgBJRwQEAgJJpiinRzaWKo4IDAACUhoADAACUhhY1AAAom6boUWsmVHAAAIDSUMEBAICSKT74U8kdDRkAAABoZAIOAABQGlrUAACgZIpi4WNVpIIDAACUhgoOAACUzCo8JVoFBwAAKA8BBwAAKA0tagAAUDarcI+aCg4AAFAaAg4AAFAaWtQAAKBkig/+VHLH5kIFBwAAKI2PXQWnpmZBUx8BoEl1XLtNUx8BoMm0b+e/gfVTk6SSvzfXVHCvpStmzZrVfE4DAACwArSoAQAApSHgAAAApSHgAAAApSHgAAAApSHgAAAApSHgAAAApSHgQD08//zzGT58eAYPHpxhw4Zl4sSJTX0kgIo74YQTMnDgwFRVVeXpp59u6uMALJaAA/VwzDHHZNSoURk/fnyOPvrojB49uqmPBFBxe+21V26++eb06tWrqY8CsEQCDizD9OnTM2HChIwcOTJJsueee2by5MmZPHlyE58MoLKGDh2anj17NvUxAJZKwIFlmDJlSrp3755WrVolSYqiSHV1dV5++eUmPhkAAB8l4EA9FEVR5+uampomOgkAAEsj4MAy9OzZM1OnTs28efOSLAw3U6ZMSXV1dROfDACAjxJwYBk6d+6cgQMHZuzYsUmSG264Ib17906fPn2a+GQAAHxUMWvWLL02sAzPPfdcjjjiiMycOTNt27bNmDFjMmDAgKY+FkBFHXfccbnpppsybdq0dOzYMWuuuWYeffTRpj4WQB0CDgAAUBpa1AAAgNIQcAAAgNIQcAAAgNIQcAAAgNIQcAAAgNIQcAAAgNIQcABW0OTJk1NVVZWqqqpceumlSZJLL7209lpTGjduXO05xo0bt8R1X//611NVVZXddtutQa9/6qmnpqqqKgMHDlzRo9b5OS7trACwNAIOUAq77bZb7S/HVVVV6dChQwYMGJCRI0fmwQcfrPh5OnXqlM022yybbbZZvZ/z32Fk8uTJK/F0AFBerZr6AACNqXXr1tl4440zZ86cPPPMM7nlllty++2355ZbbsngwYMX+5z3338/rVu3btRz7LLLLtlll10a9TUBgGVTwQFKpWvXrrntttsybty42naxefPm5aqrrkpStxXr7LPPzgYbbJCuXbvWPn/s2LHZcccd071791RXV2fffffN448/XmePu+++O1tttVW6du2az372s/nnP/+5yDmW1KJ27bXXZvjw4enZs2e6d++eoUOH5s4778ypp56aPfbYo3bdoEGDUlVVla9//etJkgULFmTMmDG1+/bp0yejRo3KpEmT6rz+ddddl0GDBqVbt27Zf//988orryz3z/L73/9+ttxyy/Tu3TudOnVK//7987WvfS2vvvrqYtf/5S9/yeabb56uXbtm+PDheeqpp+p8/9Zbb82uu+6a6urqdOvWLSNGjMjdd9+93OcDgMURcIDSqqmpWeL3/vGPf+SHP/xh2rZtm44dOyZJzj777Hz1q1/No48+mp49e6Zt27a5/fbbM2LEiNoQM23atBx44IF55pln0rJly8ycOTNf+tKX6nWeX/3qVznssMPy0EMPpWXLlunXr1/+/e9/55lnnknPnj2z/vrr164dOHBgNttss/Tr1y9Jcvzxx+fEE0/MM888k09+8pNp2bJlrr/++uyyyy6ZPn16kuTxxx/Pl7/85UyePDmtW7fOv/71r3zzm99crp9dktx2222ZOnVqqqur88lPfjLTpk3LFVdckYMOOmiRta+99loOP/zwtGzZMgsWLMhDDz2U/fffP++9916ShcHu85//fO677760b98+Xbt2zf3335+9995byAGgUQk4QKlMmzYtn/nMZ7LtttvmkEMOSZK0atUq++23X51177//fi677LI8+OCD+ec//5n33nsvp59+epLkxBNPzMMPP5wnnngim266ad5999384he/SJL87ne/y7vvvpuWLVvm9ttvz0MPPVRbZVma9957L6eddlqSZMiQIXnyySdz33335dlnn80uu+ySL37xi/n5z39eu/6SSy7JbbfdlhNOOCGTJk3KhRdemCQZM2ZM7r///jz++OPp2bNnpk2blt/+9rdJknPOOScLFixIu3bt8vDDD+fRRx/N7rvvvtw/y9/97neZNGlS7rvvvjz00EM5++yzkySPPPJIXnjhhTpr58yZk0svvTQPPPBArrjiiiTJ1KlTa//+gx/8IDU1NTnkkEPy+OOP57HHHsvuu++e+fPn59RTT13uMwLARwk4QKm8//77efjhh/PUU0+lU6dO2WWXXXLjjTcu8mb/T33qU/nsZz+bJGnZsmUmTpxYW234cDJYp06d8uijjyZJHn744STJxIkTkyTrrrtuBgwYkCTZe++9l3muiRMn5t13302SHH744WnXrl2SpG3btllnnXWW+tzHHnusthr1YYtddXV1pkyZstizbbnllunSpUuS5HOf+9wyz7YkTz75ZHbcccf07NkzVVVVOeqoo2q/99HWt6qqquy0005Jkp122qm2Ne/pp5/O66+/nhdffDHJwuDWvn37tG/fPn/5y1/qnB8AGoMhA0Cp9OrVK0888cQy130YAD703+1s66+/ftq2bVvn+x06dKizriiKxT53Zfjv1x84cGBWX331Ot/v1atXo5/t/vvvz9e//vXU1NSkQ4cOWX/99fPuu+/WturNnz+/zvr/3nNp5+/bt286deq0yJr3339/uc4JAB8l4ACrpI/+Qj5gwIB84hOfyH/+85/stNNO+fGPf1y7ZsKECZk9e3btuhtvvDHPPvts/vnPf2b99dfPDTfcsMz9+vfvnzXXXDPvvvtuLrzwwowYMSJt27bNu+++m2nTpuWTn/xk2rRpU7v+w2pSkmy66aYpiiI1NTU56KCDalviampq8sADD9SGsQEDBuSJJ57IAw88kOnTp6dz5871OtviPPzww7XB5L777ku3bt1y5pln5oc//OFi17/xxhu5/fbbs9NOO+XOO+/MrFmzkiQbbLBBOnfunF69euWll17KoEGDcsEFF6RVq4X/7+df//pXXnrppUafYgfAqkuLGkCSNm3a5Pjjj0+SnHvuudlggw2yzTbbpF+/ftl+++1zxx13JFnYXtamTZvMnz8/O+64Y4YMGZJf/vKX9Xr973znO0mSBx54IBtuuGG23nrrrLfeern55puTJP369ctqq62WZGFr2Wc+85lcf/316du3b0aNGpVk4fuDBg0alK233jp9+vTJiBEjMmHChCTJN77xjRRFkTfffDODBw/Opz/96Vx33XXL9fPYaKONav++9dZbL/M+V1999Rx88MHZcsstM3LkyCRJt27dcsABByRJTjrppCTJ9ddfn/79+2fbbbfNeuutl8022yxXXnnlcp0RABZHwAH4wLHHHpsxY8bk05/+dGbNmpUXXnghnTt3zmGHHVY7wrlbt265/PLL079//8ybNy9rrbVWzjvvvHq9/ujRo3PhhRdmyJAhmTdvXl544YX06dMn/fv3T7KwDe70009PdXV1XnvttTz88MOZNm1akuSMM87IT37yk2ywwQZ59dVX8/LLL6dXr175xje+kW222SbJwtHS559/fnr37p05c+akV69etcMRGmrHHXfMD3/4w3Tv3j2zZ8/OuuuumzPOOGOJ67t06ZLzzz+/tnVt8803z1VXXVVbldp///0zduzYDB06NLNnz86//vWvrLXWWjnggAPyxS9+cbnOCACLU8yaNWvlNo8DAABUiAoOAABQGgIOAABQGgIOAABQGgIOAABQGgIOAABQGgIOAABQGgIOAABQGgIOAABQGgIOAABQGgIOAABQGgIOAABQGgIOAABQGv8fJybqdDpQmsgAAAAASUVORK5CYII=",
      "text/plain": [
       "<Figure size 1600x800 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Training set accuracy:   0.962\n",
      "Test set accuracy:       0.961\n",
      "Training set\n",
      "Class 0: precision=0.916, recall=0.939, f1_score=0.928\n",
      "Class 1: precision=0.979, recall=0.970, f1_score=0.975\n",
      "Test set\n",
      "Class 0: precision=0.929, recall=0.904, f1_score=0.916\n",
      "Class 1: precision=0.970, recall=0.978, f1_score=0.974\n"
     ]
    }
   ],
   "source": [
    "history_model0_df = pd.DataFrame(history_model0.history)\n",
    "plot_measure(history_model0_df.loss, history_model0_df.val_loss, \"Loss\")\n",
    "plot_measure(history_model0_df.binary_accuracy, history_model0_df.val_binary_accuracy, \"Binary Accuracy\")\n",
    "\n",
    "model0.load_weights(model_file)\n",
    "predictions_train_model0 = np.rint(model0.predict(X_train))\n",
    "predictions_test_model0 = np.rint(model0.predict(X_test))\n",
    "\n",
    "plot_confusion_matrix(y_train, predictions_train_model0)\n",
    "plot_confusion_matrix(y_test, predictions_test_model0)\n",
    "\n",
    "print_accuracy_scores(y_train, y_test, predictions_train_model0, predictions_test_model0)\n",
    "print_stats(y_train, predictions_train_model0, \"Training set\")\n",
    "print_stats(y_test, predictions_test_model0, \"Test set\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 44,
   "metadata": {},
   "outputs": [],
   "source": [
    "y_train_onehot = tf.keras.utils.to_categorical(y_train, 2)\n",
    "y_valid_onehot = tf.keras.utils.to_categorical(y_valid, 2)\n",
    "y_test_onehot = tf.keras.utils.to_categorical(y_test, 2)\n",
    "\n",
    "f1_score = tf.keras.metrics.F1Score(\"weighted\")\n",
    "obj_f1_score = kt.Objective(\"val_f1_score\", \"max\")\n",
    "\n",
    "obj_accuracy = kt.Objective(\"val_accuracy\", \"max\")\n",
    "\n",
    "\n",
    "def cnn1_builder(hp):\n",
    "    index = hp.Int(\"index\", min_value=0, max_value=1, step=1)\n",
    "    activation = hp.Choice(\"activations\", activations)\n",
    "    dropout_rate = hp.Choice(\"dropout_rate\", rate)\n",
    "    l2 = hp.Choice(\"L2\", reg)\n",
    "    l1 = hp.Choice(\"L1\", reg)\n",
    "\n",
    "    init = tf.keras.initializers.HeNormal()\n",
    "\n",
    "    model1 = tf.keras.Sequential()\n",
    "\n",
    "    model1.add(tf.keras.layers.InputLayer((28, 28, 1)))\n",
    "    model1.add(tf.keras.layers.Rescaling(1.0 / 255))\n",
    "    model1.add(tf.keras.layers.GaussianNoise(0.2))\n",
    "\n",
    "    model1.add(tf.keras.layers.Conv2D(filters[index][0], 3, padding=\"same\", activation=activation))\n",
    "    model1.add(tf.keras.layers.MaxPool2D())\n",
    "\n",
    "    model1.add(tf.keras.layers.Conv2D(filters[index][1], 3, padding=\"same\", activation=activation))\n",
    "    model1.add(tf.keras.layers.MaxPool2D())\n",
    "\n",
    "    model1.add(tf.keras.layers.Conv2D(filters[index][2], 3, padding=\"same\", activation=activation))\n",
    "    model1.add(tf.keras.layers.MaxPool2D())\n",
    "\n",
    "    model1.add(tf.keras.layers.Flatten())\n",
    "\n",
    "    model1.add(\n",
    "        tf.keras.layers.Dense(\n",
    "            dense_units[index][1],\n",
    "            activation,\n",
    "            kernel_initializer=init,\n",
    "            kernel_regularizer=tf.keras.regularizers.L1L2(l1, l2),\n",
    "        ),\n",
    "    )\n",
    "    if dropout_rate > 0.0:\n",
    "        model1.add(tf.keras.layers.Dropout(dropout_rate))\n",
    "\n",
    "    model1.add(\n",
    "        tf.keras.layers.Dense(\n",
    "            dense_units[index][0],\n",
    "            activation,\n",
    "            kernel_initializer=init,\n",
    "            kernel_regularizer=tf.keras.regularizers.L1L2(l1, l2),\n",
    "        )\n",
    "    )\n",
    "    if dropout_rate > 0.0:\n",
    "        model1.add(tf.keras.layers.Dropout(dropout_rate))\n",
    "\n",
    "    model1.add(\n",
    "        tf.keras.layers.Dense(\n",
    "            2,\n",
    "            \"softmax\",\n",
    "            kernel_initializer=init,\n",
    "            kernel_regularizer=tf.keras.regularizers.L1L2(l1, l2),\n",
    "        )\n",
    "    )\n",
    "\n",
    "    model1.compile(optimizer=\"adam\", loss=\"categorical_crossentropy\", metrics=[\"accuracy\", f1_score])\n",
    "\n",
    "    return model1"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 45,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Reloading Tuner from ./tuners/cnn1_weighted/tuner0.json\n"
     ]
    }
   ],
   "source": [
    "early_stop = tf.keras.callbacks.EarlyStopping(monitor=\"loss\", mode=\"min\", patience=3)\n",
    "\n",
    "tuner = kt.Hyperband(cnn1_builder, obj_accuracy, seed=42, project_name=\"tuners/cnn1_weighted\")\n",
    "tuner.search(\n",
    "    X_train,\n",
    "    y_train_onehot,\n",
    "    validation_data=(X_valid, y_valid_onehot),\n",
    "    callbacks=[early_stop],\n",
    "    class_weight=class_weights,\n",
    ")\n",
    "best_hps = tuner.get_best_hyperparameters(num_trials=1)[0]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 46,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Best Filter:  [24, 48, 96]\n",
      "Best Dense units:  [256, 512]\n",
      "Best Activation Function:  leaky_relu\n",
      "Best L2 value:  0.001\n",
      "Best Dropout Rate:  0.5\n"
     ]
    }
   ],
   "source": [
    "print(\"Best Filter: \", filters[best_hps.get(\"index\")])\n",
    "print(\"Best Dense units: \", dense_units[best_hps.get(\"index\")])\n",
    "print(\"Best Activation Function: \", best_hps.get(\"activations\"))\n",
    "\n",
    "if best_hps.get(\"L1\") > 0:\n",
    "    print(\"Best L1 value: \", best_hps.get(\"L1\"))\n",
    "\n",
    "if best_hps.get(\"L2\") > 0:\n",
    "    print(\"Best L2 value: \", best_hps.get(\"L2\"))\n",
    "\n",
    "print(\"Best Dropout Rate: \", best_hps.get(\"dropout_rate\"))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 47,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<pre style=\"white-space:pre;overflow-x:auto;line-height:normal;font-family:Menlo,'DejaVu Sans Mono',consolas,'Courier New',monospace\"><span style=\"font-weight: bold\">Model: \"sequential_1\"</span>\n",
       "</pre>\n"
      ],
      "text/plain": [
       "\u001b[1mModel: \"sequential_1\"\u001b[0m\n"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre style=\"white-space:pre;overflow-x:auto;line-height:normal;font-family:Menlo,'DejaVu Sans Mono',consolas,'Courier New',monospace\">┏━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━┓\n",
       "┃<span style=\"font-weight: bold\"> Layer (type)                    </span>┃<span style=\"font-weight: bold\"> Output Shape           </span>┃<span style=\"font-weight: bold\">       Param # </span>┃\n",
       "┡━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━┩\n",
       "│ rescaling_1 (<span style=\"color: #0087ff; text-decoration-color: #0087ff\">Rescaling</span>)         │ (<span style=\"color: #00d7ff; text-decoration-color: #00d7ff\">None</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">28</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">28</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">1</span>)      │             <span style=\"color: #00af00; text-decoration-color: #00af00\">0</span> │\n",
       "├─────────────────────────────────┼────────────────────────┼───────────────┤\n",
       "│ gaussian_noise_1                │ (<span style=\"color: #00d7ff; text-decoration-color: #00d7ff\">None</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">28</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">28</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">1</span>)      │             <span style=\"color: #00af00; text-decoration-color: #00af00\">0</span> │\n",
       "│ (<span style=\"color: #0087ff; text-decoration-color: #0087ff\">GaussianNoise</span>)                 │                        │               │\n",
       "├─────────────────────────────────┼────────────────────────┼───────────────┤\n",
       "│ conv2d_3 (<span style=\"color: #0087ff; text-decoration-color: #0087ff\">Conv2D</span>)               │ (<span style=\"color: #00d7ff; text-decoration-color: #00d7ff\">None</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">28</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">28</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">24</span>)     │           <span style=\"color: #00af00; text-decoration-color: #00af00\">240</span> │\n",
       "├─────────────────────────────────┼────────────────────────┼───────────────┤\n",
       "│ max_pooling2d_3 (<span style=\"color: #0087ff; text-decoration-color: #0087ff\">MaxPooling2D</span>)  │ (<span style=\"color: #00d7ff; text-decoration-color: #00d7ff\">None</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">14</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">14</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">24</span>)     │             <span style=\"color: #00af00; text-decoration-color: #00af00\">0</span> │\n",
       "├─────────────────────────────────┼────────────────────────┼───────────────┤\n",
       "│ conv2d_4 (<span style=\"color: #0087ff; text-decoration-color: #0087ff\">Conv2D</span>)               │ (<span style=\"color: #00d7ff; text-decoration-color: #00d7ff\">None</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">14</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">14</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">48</span>)     │        <span style=\"color: #00af00; text-decoration-color: #00af00\">10,416</span> │\n",
       "├─────────────────────────────────┼────────────────────────┼───────────────┤\n",
       "│ max_pooling2d_4 (<span style=\"color: #0087ff; text-decoration-color: #0087ff\">MaxPooling2D</span>)  │ (<span style=\"color: #00d7ff; text-decoration-color: #00d7ff\">None</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">7</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">7</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">48</span>)       │             <span style=\"color: #00af00; text-decoration-color: #00af00\">0</span> │\n",
       "├─────────────────────────────────┼────────────────────────┼───────────────┤\n",
       "│ conv2d_5 (<span style=\"color: #0087ff; text-decoration-color: #0087ff\">Conv2D</span>)               │ (<span style=\"color: #00d7ff; text-decoration-color: #00d7ff\">None</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">7</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">7</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">96</span>)       │        <span style=\"color: #00af00; text-decoration-color: #00af00\">41,568</span> │\n",
       "├─────────────────────────────────┼────────────────────────┼───────────────┤\n",
       "│ max_pooling2d_5 (<span style=\"color: #0087ff; text-decoration-color: #0087ff\">MaxPooling2D</span>)  │ (<span style=\"color: #00d7ff; text-decoration-color: #00d7ff\">None</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">3</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">3</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">96</span>)       │             <span style=\"color: #00af00; text-decoration-color: #00af00\">0</span> │\n",
       "├─────────────────────────────────┼────────────────────────┼───────────────┤\n",
       "│ flatten_1 (<span style=\"color: #0087ff; text-decoration-color: #0087ff\">Flatten</span>)             │ (<span style=\"color: #00d7ff; text-decoration-color: #00d7ff\">None</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">864</span>)            │             <span style=\"color: #00af00; text-decoration-color: #00af00\">0</span> │\n",
       "├─────────────────────────────────┼────────────────────────┼───────────────┤\n",
       "│ dense_3 (<span style=\"color: #0087ff; text-decoration-color: #0087ff\">Dense</span>)                 │ (<span style=\"color: #00d7ff; text-decoration-color: #00d7ff\">None</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">512</span>)            │       <span style=\"color: #00af00; text-decoration-color: #00af00\">442,880</span> │\n",
       "├─────────────────────────────────┼────────────────────────┼───────────────┤\n",
       "│ dropout_2 (<span style=\"color: #0087ff; text-decoration-color: #0087ff\">Dropout</span>)             │ (<span style=\"color: #00d7ff; text-decoration-color: #00d7ff\">None</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">512</span>)            │             <span style=\"color: #00af00; text-decoration-color: #00af00\">0</span> │\n",
       "├─────────────────────────────────┼────────────────────────┼───────────────┤\n",
       "│ dense_4 (<span style=\"color: #0087ff; text-decoration-color: #0087ff\">Dense</span>)                 │ (<span style=\"color: #00d7ff; text-decoration-color: #00d7ff\">None</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">256</span>)            │       <span style=\"color: #00af00; text-decoration-color: #00af00\">131,328</span> │\n",
       "├─────────────────────────────────┼────────────────────────┼───────────────┤\n",
       "│ dropout_3 (<span style=\"color: #0087ff; text-decoration-color: #0087ff\">Dropout</span>)             │ (<span style=\"color: #00d7ff; text-decoration-color: #00d7ff\">None</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">256</span>)            │             <span style=\"color: #00af00; text-decoration-color: #00af00\">0</span> │\n",
       "├─────────────────────────────────┼────────────────────────┼───────────────┤\n",
       "│ dense_5 (<span style=\"color: #0087ff; text-decoration-color: #0087ff\">Dense</span>)                 │ (<span style=\"color: #00d7ff; text-decoration-color: #00d7ff\">None</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">2</span>)              │           <span style=\"color: #00af00; text-decoration-color: #00af00\">514</span> │\n",
       "└─────────────────────────────────┴────────────────────────┴───────────────┘\n",
       "</pre>\n"
      ],
      "text/plain": [
       "┏━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━┓\n",
       "┃\u001b[1m \u001b[0m\u001b[1mLayer (type)                   \u001b[0m\u001b[1m \u001b[0m┃\u001b[1m \u001b[0m\u001b[1mOutput Shape          \u001b[0m\u001b[1m \u001b[0m┃\u001b[1m \u001b[0m\u001b[1m      Param #\u001b[0m\u001b[1m \u001b[0m┃\n",
       "┡━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━┩\n",
       "│ rescaling_1 (\u001b[38;5;33mRescaling\u001b[0m)         │ (\u001b[38;5;45mNone\u001b[0m, \u001b[38;5;34m28\u001b[0m, \u001b[38;5;34m28\u001b[0m, \u001b[38;5;34m1\u001b[0m)      │             \u001b[38;5;34m0\u001b[0m │\n",
       "├─────────────────────────────────┼────────────────────────┼───────────────┤\n",
       "│ gaussian_noise_1                │ (\u001b[38;5;45mNone\u001b[0m, \u001b[38;5;34m28\u001b[0m, \u001b[38;5;34m28\u001b[0m, \u001b[38;5;34m1\u001b[0m)      │             \u001b[38;5;34m0\u001b[0m │\n",
       "│ (\u001b[38;5;33mGaussianNoise\u001b[0m)                 │                        │               │\n",
       "├─────────────────────────────────┼────────────────────────┼───────────────┤\n",
       "│ conv2d_3 (\u001b[38;5;33mConv2D\u001b[0m)               │ (\u001b[38;5;45mNone\u001b[0m, \u001b[38;5;34m28\u001b[0m, \u001b[38;5;34m28\u001b[0m, \u001b[38;5;34m24\u001b[0m)     │           \u001b[38;5;34m240\u001b[0m │\n",
       "├─────────────────────────────────┼────────────────────────┼───────────────┤\n",
       "│ max_pooling2d_3 (\u001b[38;5;33mMaxPooling2D\u001b[0m)  │ (\u001b[38;5;45mNone\u001b[0m, \u001b[38;5;34m14\u001b[0m, \u001b[38;5;34m14\u001b[0m, \u001b[38;5;34m24\u001b[0m)     │             \u001b[38;5;34m0\u001b[0m │\n",
       "├─────────────────────────────────┼────────────────────────┼───────────────┤\n",
       "│ conv2d_4 (\u001b[38;5;33mConv2D\u001b[0m)               │ (\u001b[38;5;45mNone\u001b[0m, \u001b[38;5;34m14\u001b[0m, \u001b[38;5;34m14\u001b[0m, \u001b[38;5;34m48\u001b[0m)     │        \u001b[38;5;34m10,416\u001b[0m │\n",
       "├─────────────────────────────────┼────────────────────────┼───────────────┤\n",
       "│ max_pooling2d_4 (\u001b[38;5;33mMaxPooling2D\u001b[0m)  │ (\u001b[38;5;45mNone\u001b[0m, \u001b[38;5;34m7\u001b[0m, \u001b[38;5;34m7\u001b[0m, \u001b[38;5;34m48\u001b[0m)       │             \u001b[38;5;34m0\u001b[0m │\n",
       "├─────────────────────────────────┼────────────────────────┼───────────────┤\n",
       "│ conv2d_5 (\u001b[38;5;33mConv2D\u001b[0m)               │ (\u001b[38;5;45mNone\u001b[0m, \u001b[38;5;34m7\u001b[0m, \u001b[38;5;34m7\u001b[0m, \u001b[38;5;34m96\u001b[0m)       │        \u001b[38;5;34m41,568\u001b[0m │\n",
       "├─────────────────────────────────┼────────────────────────┼───────────────┤\n",
       "│ max_pooling2d_5 (\u001b[38;5;33mMaxPooling2D\u001b[0m)  │ (\u001b[38;5;45mNone\u001b[0m, \u001b[38;5;34m3\u001b[0m, \u001b[38;5;34m3\u001b[0m, \u001b[38;5;34m96\u001b[0m)       │             \u001b[38;5;34m0\u001b[0m │\n",
       "├─────────────────────────────────┼────────────────────────┼───────────────┤\n",
       "│ flatten_1 (\u001b[38;5;33mFlatten\u001b[0m)             │ (\u001b[38;5;45mNone\u001b[0m, \u001b[38;5;34m864\u001b[0m)            │             \u001b[38;5;34m0\u001b[0m │\n",
       "├─────────────────────────────────┼────────────────────────┼───────────────┤\n",
       "│ dense_3 (\u001b[38;5;33mDense\u001b[0m)                 │ (\u001b[38;5;45mNone\u001b[0m, \u001b[38;5;34m512\u001b[0m)            │       \u001b[38;5;34m442,880\u001b[0m │\n",
       "├─────────────────────────────────┼────────────────────────┼───────────────┤\n",
       "│ dropout_2 (\u001b[38;5;33mDropout\u001b[0m)             │ (\u001b[38;5;45mNone\u001b[0m, \u001b[38;5;34m512\u001b[0m)            │             \u001b[38;5;34m0\u001b[0m │\n",
       "├─────────────────────────────────┼────────────────────────┼───────────────┤\n",
       "│ dense_4 (\u001b[38;5;33mDense\u001b[0m)                 │ (\u001b[38;5;45mNone\u001b[0m, \u001b[38;5;34m256\u001b[0m)            │       \u001b[38;5;34m131,328\u001b[0m │\n",
       "├─────────────────────────────────┼────────────────────────┼───────────────┤\n",
       "│ dropout_3 (\u001b[38;5;33mDropout\u001b[0m)             │ (\u001b[38;5;45mNone\u001b[0m, \u001b[38;5;34m256\u001b[0m)            │             \u001b[38;5;34m0\u001b[0m │\n",
       "├─────────────────────────────────┼────────────────────────┼───────────────┤\n",
       "│ dense_5 (\u001b[38;5;33mDense\u001b[0m)                 │ (\u001b[38;5;45mNone\u001b[0m, \u001b[38;5;34m2\u001b[0m)              │           \u001b[38;5;34m514\u001b[0m │\n",
       "└─────────────────────────────────┴────────────────────────┴───────────────┘\n"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre style=\"white-space:pre;overflow-x:auto;line-height:normal;font-family:Menlo,'DejaVu Sans Mono',consolas,'Courier New',monospace\"><span style=\"font-weight: bold\"> Total params: </span><span style=\"color: #00af00; text-decoration-color: #00af00\">626,946</span> (2.39 MB)\n",
       "</pre>\n"
      ],
      "text/plain": [
       "\u001b[1m Total params: \u001b[0m\u001b[38;5;34m626,946\u001b[0m (2.39 MB)\n"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre style=\"white-space:pre;overflow-x:auto;line-height:normal;font-family:Menlo,'DejaVu Sans Mono',consolas,'Courier New',monospace\"><span style=\"font-weight: bold\"> Trainable params: </span><span style=\"color: #00af00; text-decoration-color: #00af00\">626,946</span> (2.39 MB)\n",
       "</pre>\n"
      ],
      "text/plain": [
       "\u001b[1m Trainable params: \u001b[0m\u001b[38;5;34m626,946\u001b[0m (2.39 MB)\n"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre style=\"white-space:pre;overflow-x:auto;line-height:normal;font-family:Menlo,'DejaVu Sans Mono',consolas,'Courier New',monospace\"><span style=\"font-weight: bold\"> Non-trainable params: </span><span style=\"color: #00af00; text-decoration-color: #00af00\">0</span> (0.00 B)\n",
       "</pre>\n"
      ],
      "text/plain": [
       "\u001b[1m Non-trainable params: \u001b[0m\u001b[38;5;34m0\u001b[0m (0.00 B)\n"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "model1 = tuner.hypermodel.build(best_hps)\n",
    "model1.summary()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 48,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/100\n",
      "\u001b[1m115/115\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 67ms/step - accuracy: 0.5609 - f1_score: 0.5866 - loss: 1.7032\n",
      "Epoch 1: val_accuracy improved from -inf to 0.90955, saving model to ./best_model/cnn1_weighted.weights.h5\n",
      "\u001b[1m115/115\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m13s\u001b[0m 81ms/step - accuracy: 0.5619 - f1_score: 0.5876 - loss: 1.6995 - val_accuracy: 0.9096 - val_f1_score: 0.9067 - val_loss: 0.6012\n",
      "Epoch 2/100\n",
      "\u001b[1m115/115\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - accuracy: 0.8898 - f1_score: 0.8933 - loss: 0.5723\n",
      "Epoch 2: val_accuracy did not improve from 0.90955\n",
      "\u001b[1m115/115\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m1s\u001b[0m 6ms/step - accuracy: 0.8899 - f1_score: 0.8933 - loss: 0.5718 - val_accuracy: 0.8815 - val_f1_score: 0.8724 - val_loss: 0.4456\n",
      "Epoch 3/100\n",
      "\u001b[1m107/115\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m━━\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - accuracy: 0.8945 - f1_score: 0.8978 - loss: 0.4195\n",
      "Epoch 3: val_accuracy did not improve from 0.90955\n",
      "\u001b[1m115/115\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m1s\u001b[0m 5ms/step - accuracy: 0.8948 - f1_score: 0.8982 - loss: 0.4173 - val_accuracy: 0.9070 - val_f1_score: 0.9027 - val_loss: 0.3482\n",
      "Epoch 4/100\n",
      "\u001b[1m112/115\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m━\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - accuracy: 0.9158 - f1_score: 0.9181 - loss: 0.3257\n",
      "Epoch 4: val_accuracy improved from 0.90955 to 0.94140, saving model to ./best_model/cnn1_weighted.weights.h5\n",
      "\u001b[1m115/115\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m1s\u001b[0m 5ms/step - accuracy: 0.9156 - f1_score: 0.9179 - loss: 0.3257 - val_accuracy: 0.9414 - val_f1_score: 0.9409 - val_loss: 0.2556\n",
      "Epoch 5/100\n",
      "\u001b[1m113/115\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m━\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - accuracy: 0.9106 - f1_score: 0.9135 - loss: 0.2944\n",
      "Epoch 5: val_accuracy did not improve from 0.94140\n",
      "\u001b[1m115/115\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m1s\u001b[0m 5ms/step - accuracy: 0.9107 - f1_score: 0.9136 - loss: 0.2941 - val_accuracy: 0.9236 - val_f1_score: 0.9209 - val_loss: 0.2779\n",
      "Epoch 6/100\n",
      "\u001b[1m107/115\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m━━\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - accuracy: 0.9280 - f1_score: 0.9298 - loss: 0.2494\n",
      "Epoch 6: val_accuracy did not improve from 0.94140\n",
      "\u001b[1m115/115\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m1s\u001b[0m 5ms/step - accuracy: 0.9278 - f1_score: 0.9296 - loss: 0.2500 - val_accuracy: 0.9325 - val_f1_score: 0.9309 - val_loss: 0.2461\n",
      "Epoch 7/100\n",
      "\u001b[1m113/115\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m━\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - accuracy: 0.9213 - f1_score: 0.9235 - loss: 0.2370\n",
      "Epoch 7: val_accuracy did not improve from 0.94140\n",
      "\u001b[1m115/115\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m1s\u001b[0m 5ms/step - accuracy: 0.9212 - f1_score: 0.9234 - loss: 0.2371 - val_accuracy: 0.9274 - val_f1_score: 0.9253 - val_loss: 0.2411\n",
      "Epoch 8/100\n",
      "\u001b[1m111/115\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m━\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - accuracy: 0.9296 - f1_score: 0.9313 - loss: 0.2289\n",
      "Epoch 8: val_accuracy did not improve from 0.94140\n",
      "\u001b[1m115/115\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m1s\u001b[0m 5ms/step - accuracy: 0.9294 - f1_score: 0.9311 - loss: 0.2289 - val_accuracy: 0.9223 - val_f1_score: 0.9195 - val_loss: 0.2341\n",
      "Epoch 9/100\n",
      "\u001b[1m115/115\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - accuracy: 0.9275 - f1_score: 0.9293 - loss: 0.2207\n",
      "Epoch 9: val_accuracy improved from 0.94140 to 0.94522, saving model to ./best_model/cnn1_weighted.weights.h5\n",
      "\u001b[1m115/115\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m1s\u001b[0m 6ms/step - accuracy: 0.9274 - f1_score: 0.9293 - loss: 0.2207 - val_accuracy: 0.9452 - val_f1_score: 0.9443 - val_loss: 0.2018\n",
      "Epoch 10/100\n",
      "\u001b[1m109/115\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m━━\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - accuracy: 0.9281 - f1_score: 0.9299 - loss: 0.2127\n",
      "Epoch 10: val_accuracy did not improve from 0.94522\n",
      "\u001b[1m115/115\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m1s\u001b[0m 5ms/step - accuracy: 0.9282 - f1_score: 0.9300 - loss: 0.2122 - val_accuracy: 0.9376 - val_f1_score: 0.9363 - val_loss: 0.1908\n",
      "Epoch 11/100\n",
      "\u001b[1m112/115\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m━\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - accuracy: 0.9342 - f1_score: 0.9357 - loss: 0.1987\n",
      "Epoch 11: val_accuracy did not improve from 0.94522\n",
      "\u001b[1m115/115\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m1s\u001b[0m 5ms/step - accuracy: 0.9339 - f1_score: 0.9355 - loss: 0.1993 - val_accuracy: 0.9248 - val_f1_score: 0.9223 - val_loss: 0.2144\n",
      "Epoch 12/100\n",
      "\u001b[1m110/115\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m━\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - accuracy: 0.9427 - f1_score: 0.9440 - loss: 0.1875\n",
      "Epoch 12: val_accuracy improved from 0.94522 to 0.94777, saving model to ./best_model/cnn1_weighted.weights.h5\n",
      "\u001b[1m115/115\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m1s\u001b[0m 5ms/step - accuracy: 0.9423 - f1_score: 0.9436 - loss: 0.1878 - val_accuracy: 0.9478 - val_f1_score: 0.9477 - val_loss: 0.1733\n",
      "Epoch 13/100\n",
      "\u001b[1m111/115\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m━\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - accuracy: 0.9299 - f1_score: 0.9318 - loss: 0.1965\n",
      "Epoch 13: val_accuracy improved from 0.94777 to 0.95032, saving model to ./best_model/cnn1_weighted.weights.h5\n",
      "\u001b[1m115/115\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m1s\u001b[0m 5ms/step - accuracy: 0.9295 - f1_score: 0.9314 - loss: 0.1974 - val_accuracy: 0.9503 - val_f1_score: 0.9504 - val_loss: 0.1807\n",
      "Epoch 14/100\n",
      "\u001b[1m110/115\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m━\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - accuracy: 0.9428 - f1_score: 0.9440 - loss: 0.1894\n",
      "Epoch 14: val_accuracy did not improve from 0.95032\n",
      "\u001b[1m115/115\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m1s\u001b[0m 5ms/step - accuracy: 0.9424 - f1_score: 0.9436 - loss: 0.1898 - val_accuracy: 0.9503 - val_f1_score: 0.9504 - val_loss: 0.1654\n",
      "Epoch 15/100\n",
      "\u001b[1m111/115\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m━\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - accuracy: 0.9375 - f1_score: 0.9390 - loss: 0.1840\n",
      "Epoch 15: val_accuracy did not improve from 0.95032\n",
      "\u001b[1m115/115\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m1s\u001b[0m 5ms/step - accuracy: 0.9373 - f1_score: 0.9388 - loss: 0.1845 - val_accuracy: 0.9465 - val_f1_score: 0.9453 - val_loss: 0.1814\n"
     ]
    }
   ],
   "source": [
    "model_file = \"./best_model/cnn1_weighted.weights.h5\"\n",
    "checkpoint = tf.keras.callbacks.ModelCheckpoint(model_file, \"val_accuracy\", 1, True, True, \"max\")\n",
    "history_model1 = model1.fit(\n",
    "    X_train,\n",
    "    y_train_onehot,\n",
    "    epochs=100,\n",
    "    validation_data=(X_valid, y_valid_onehot),\n",
    "    callbacks=[early_stop, checkpoint],\n",
    "    class_weight=class_weights,\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 49,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 1200x800 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 1200x800 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\u001b[1m115/115\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 2ms/step\n",
      "\u001b[1m25/25\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "<Figure size 1200x600 with 0 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 1600x800 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "<Figure size 1200x600 with 0 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 1600x800 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Training set accuracy:   0.958\n",
      "Test set accuracy:       0.950\n",
      "Training set\n",
      "Class 0: precision=0.900, recall=0.944, f1_score=0.921\n",
      "Class 1: precision=0.980, recall=0.964, f1_score=0.972\n",
      "Test set\n",
      "Class 0: precision=0.889, recall=0.904, f1_score=0.897\n",
      "Class 1: precision=0.970, recall=0.965, f1_score=0.967\n"
     ]
    }
   ],
   "source": [
    "history_model1_df = pd.DataFrame(history_model1.history)\n",
    "plot_measure(history_model1_df.loss, history_model1_df.val_loss, \"Loss\")\n",
    "plot_measure(history_model1_df.accuracy, history_model1_df.val_accuracy, \"Accuracy\")\n",
    "\n",
    "model1.load_weights(model_file)\n",
    "predictions_train_model1 = np.argmax(model1.predict(X_train), axis=-1)\n",
    "predictions_test_model1 = np.argmax(model1.predict(X_test), axis=-1)\n",
    "\n",
    "plot_confusion_matrix(y_train, predictions_train_model1)\n",
    "plot_confusion_matrix(y_test, predictions_test_model1)\n",
    "\n",
    "print_accuracy_scores(y_train, y_test, predictions_train_model1, predictions_test_model1)\n",
    "print_stats(y_train, predictions_train_model1, \"Training set\")\n",
    "print_stats(y_test, predictions_test_model1, \"Test set\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 50,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Reloading Tuner from ./tuners/cnn1_weighted_f1/tuner0.json\n"
     ]
    }
   ],
   "source": [
    "early_stop = tf.keras.callbacks.EarlyStopping(monitor=\"loss\", mode=\"min\", patience=5)\n",
    "\n",
    "tuner = kt.Hyperband(cnn1_builder, obj_f1_score, seed=42, project_name=\"tuners/cnn1_weighted_f1\")\n",
    "tuner.search(\n",
    "    X_train,\n",
    "    y_train_onehot,\n",
    "    validation_data=(X_valid, y_valid_onehot),\n",
    "    callbacks=[early_stop],\n",
    "    class_weight=class_weights,\n",
    ")\n",
    "best_hps = tuner.get_best_hyperparameters(num_trials=1)[0]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 51,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Best Filter:  [24, 48, 96]\n",
      "Best Dense units:  [256, 512]\n",
      "Best Activations Functions:  leaky_relu\n",
      "Best L2 value:  0.001\n",
      "Best Dropout Rate:  0.5\n"
     ]
    }
   ],
   "source": [
    "print(\"Best Filter: \", filters[best_hps.get(\"index\")])\n",
    "print(\"Best Dense units: \", dense_units[best_hps.get(\"index\")])\n",
    "print(\"Best Activations Functions: \", best_hps.get(\"activations\"))\n",
    "\n",
    "if best_hps.get(\"L1\") > 0:\n",
    "    print(\"Best L1 value: \", best_hps.get(\"L1\"))\n",
    "\n",
    "if best_hps.get(\"L2\") > 0:\n",
    "    print(\"Best L2 value: \", best_hps.get(\"L2\"))\n",
    "\n",
    "print(\"Best Dropout Rate: \", best_hps.get(\"dropout_rate\"))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 52,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<pre style=\"white-space:pre;overflow-x:auto;line-height:normal;font-family:Menlo,'DejaVu Sans Mono',consolas,'Courier New',monospace\"><span style=\"font-weight: bold\">Model: \"sequential_2\"</span>\n",
       "</pre>\n"
      ],
      "text/plain": [
       "\u001b[1mModel: \"sequential_2\"\u001b[0m\n"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre style=\"white-space:pre;overflow-x:auto;line-height:normal;font-family:Menlo,'DejaVu Sans Mono',consolas,'Courier New',monospace\">┏━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━┓\n",
       "┃<span style=\"font-weight: bold\"> Layer (type)                    </span>┃<span style=\"font-weight: bold\"> Output Shape           </span>┃<span style=\"font-weight: bold\">       Param # </span>┃\n",
       "┡━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━┩\n",
       "│ rescaling_2 (<span style=\"color: #0087ff; text-decoration-color: #0087ff\">Rescaling</span>)         │ (<span style=\"color: #00d7ff; text-decoration-color: #00d7ff\">None</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">28</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">28</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">1</span>)      │             <span style=\"color: #00af00; text-decoration-color: #00af00\">0</span> │\n",
       "├─────────────────────────────────┼────────────────────────┼───────────────┤\n",
       "│ gaussian_noise_2                │ (<span style=\"color: #00d7ff; text-decoration-color: #00d7ff\">None</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">28</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">28</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">1</span>)      │             <span style=\"color: #00af00; text-decoration-color: #00af00\">0</span> │\n",
       "│ (<span style=\"color: #0087ff; text-decoration-color: #0087ff\">GaussianNoise</span>)                 │                        │               │\n",
       "├─────────────────────────────────┼────────────────────────┼───────────────┤\n",
       "│ conv2d_6 (<span style=\"color: #0087ff; text-decoration-color: #0087ff\">Conv2D</span>)               │ (<span style=\"color: #00d7ff; text-decoration-color: #00d7ff\">None</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">28</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">28</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">24</span>)     │           <span style=\"color: #00af00; text-decoration-color: #00af00\">240</span> │\n",
       "├─────────────────────────────────┼────────────────────────┼───────────────┤\n",
       "│ max_pooling2d_6 (<span style=\"color: #0087ff; text-decoration-color: #0087ff\">MaxPooling2D</span>)  │ (<span style=\"color: #00d7ff; text-decoration-color: #00d7ff\">None</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">14</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">14</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">24</span>)     │             <span style=\"color: #00af00; text-decoration-color: #00af00\">0</span> │\n",
       "├─────────────────────────────────┼────────────────────────┼───────────────┤\n",
       "│ conv2d_7 (<span style=\"color: #0087ff; text-decoration-color: #0087ff\">Conv2D</span>)               │ (<span style=\"color: #00d7ff; text-decoration-color: #00d7ff\">None</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">14</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">14</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">48</span>)     │        <span style=\"color: #00af00; text-decoration-color: #00af00\">10,416</span> │\n",
       "├─────────────────────────────────┼────────────────────────┼───────────────┤\n",
       "│ max_pooling2d_7 (<span style=\"color: #0087ff; text-decoration-color: #0087ff\">MaxPooling2D</span>)  │ (<span style=\"color: #00d7ff; text-decoration-color: #00d7ff\">None</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">7</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">7</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">48</span>)       │             <span style=\"color: #00af00; text-decoration-color: #00af00\">0</span> │\n",
       "├─────────────────────────────────┼────────────────────────┼───────────────┤\n",
       "│ conv2d_8 (<span style=\"color: #0087ff; text-decoration-color: #0087ff\">Conv2D</span>)               │ (<span style=\"color: #00d7ff; text-decoration-color: #00d7ff\">None</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">7</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">7</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">96</span>)       │        <span style=\"color: #00af00; text-decoration-color: #00af00\">41,568</span> │\n",
       "├─────────────────────────────────┼────────────────────────┼───────────────┤\n",
       "│ max_pooling2d_8 (<span style=\"color: #0087ff; text-decoration-color: #0087ff\">MaxPooling2D</span>)  │ (<span style=\"color: #00d7ff; text-decoration-color: #00d7ff\">None</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">3</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">3</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">96</span>)       │             <span style=\"color: #00af00; text-decoration-color: #00af00\">0</span> │\n",
       "├─────────────────────────────────┼────────────────────────┼───────────────┤\n",
       "│ flatten_2 (<span style=\"color: #0087ff; text-decoration-color: #0087ff\">Flatten</span>)             │ (<span style=\"color: #00d7ff; text-decoration-color: #00d7ff\">None</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">864</span>)            │             <span style=\"color: #00af00; text-decoration-color: #00af00\">0</span> │\n",
       "├─────────────────────────────────┼────────────────────────┼───────────────┤\n",
       "│ dense_6 (<span style=\"color: #0087ff; text-decoration-color: #0087ff\">Dense</span>)                 │ (<span style=\"color: #00d7ff; text-decoration-color: #00d7ff\">None</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">512</span>)            │       <span style=\"color: #00af00; text-decoration-color: #00af00\">442,880</span> │\n",
       "├─────────────────────────────────┼────────────────────────┼───────────────┤\n",
       "│ dropout_4 (<span style=\"color: #0087ff; text-decoration-color: #0087ff\">Dropout</span>)             │ (<span style=\"color: #00d7ff; text-decoration-color: #00d7ff\">None</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">512</span>)            │             <span style=\"color: #00af00; text-decoration-color: #00af00\">0</span> │\n",
       "├─────────────────────────────────┼────────────────────────┼───────────────┤\n",
       "│ dense_7 (<span style=\"color: #0087ff; text-decoration-color: #0087ff\">Dense</span>)                 │ (<span style=\"color: #00d7ff; text-decoration-color: #00d7ff\">None</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">256</span>)            │       <span style=\"color: #00af00; text-decoration-color: #00af00\">131,328</span> │\n",
       "├─────────────────────────────────┼────────────────────────┼───────────────┤\n",
       "│ dropout_5 (<span style=\"color: #0087ff; text-decoration-color: #0087ff\">Dropout</span>)             │ (<span style=\"color: #00d7ff; text-decoration-color: #00d7ff\">None</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">256</span>)            │             <span style=\"color: #00af00; text-decoration-color: #00af00\">0</span> │\n",
       "├─────────────────────────────────┼────────────────────────┼───────────────┤\n",
       "│ dense_8 (<span style=\"color: #0087ff; text-decoration-color: #0087ff\">Dense</span>)                 │ (<span style=\"color: #00d7ff; text-decoration-color: #00d7ff\">None</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">2</span>)              │           <span style=\"color: #00af00; text-decoration-color: #00af00\">514</span> │\n",
       "└─────────────────────────────────┴────────────────────────┴───────────────┘\n",
       "</pre>\n"
      ],
      "text/plain": [
       "┏━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━┓\n",
       "┃\u001b[1m \u001b[0m\u001b[1mLayer (type)                   \u001b[0m\u001b[1m \u001b[0m┃\u001b[1m \u001b[0m\u001b[1mOutput Shape          \u001b[0m\u001b[1m \u001b[0m┃\u001b[1m \u001b[0m\u001b[1m      Param #\u001b[0m\u001b[1m \u001b[0m┃\n",
       "┡━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━┩\n",
       "│ rescaling_2 (\u001b[38;5;33mRescaling\u001b[0m)         │ (\u001b[38;5;45mNone\u001b[0m, \u001b[38;5;34m28\u001b[0m, \u001b[38;5;34m28\u001b[0m, \u001b[38;5;34m1\u001b[0m)      │             \u001b[38;5;34m0\u001b[0m │\n",
       "├─────────────────────────────────┼────────────────────────┼───────────────┤\n",
       "│ gaussian_noise_2                │ (\u001b[38;5;45mNone\u001b[0m, \u001b[38;5;34m28\u001b[0m, \u001b[38;5;34m28\u001b[0m, \u001b[38;5;34m1\u001b[0m)      │             \u001b[38;5;34m0\u001b[0m │\n",
       "│ (\u001b[38;5;33mGaussianNoise\u001b[0m)                 │                        │               │\n",
       "├─────────────────────────────────┼────────────────────────┼───────────────┤\n",
       "│ conv2d_6 (\u001b[38;5;33mConv2D\u001b[0m)               │ (\u001b[38;5;45mNone\u001b[0m, \u001b[38;5;34m28\u001b[0m, \u001b[38;5;34m28\u001b[0m, \u001b[38;5;34m24\u001b[0m)     │           \u001b[38;5;34m240\u001b[0m │\n",
       "├─────────────────────────────────┼────────────────────────┼───────────────┤\n",
       "│ max_pooling2d_6 (\u001b[38;5;33mMaxPooling2D\u001b[0m)  │ (\u001b[38;5;45mNone\u001b[0m, \u001b[38;5;34m14\u001b[0m, \u001b[38;5;34m14\u001b[0m, \u001b[38;5;34m24\u001b[0m)     │             \u001b[38;5;34m0\u001b[0m │\n",
       "├─────────────────────────────────┼────────────────────────┼───────────────┤\n",
       "│ conv2d_7 (\u001b[38;5;33mConv2D\u001b[0m)               │ (\u001b[38;5;45mNone\u001b[0m, \u001b[38;5;34m14\u001b[0m, \u001b[38;5;34m14\u001b[0m, \u001b[38;5;34m48\u001b[0m)     │        \u001b[38;5;34m10,416\u001b[0m │\n",
       "├─────────────────────────────────┼────────────────────────┼───────────────┤\n",
       "│ max_pooling2d_7 (\u001b[38;5;33mMaxPooling2D\u001b[0m)  │ (\u001b[38;5;45mNone\u001b[0m, \u001b[38;5;34m7\u001b[0m, \u001b[38;5;34m7\u001b[0m, \u001b[38;5;34m48\u001b[0m)       │             \u001b[38;5;34m0\u001b[0m │\n",
       "├─────────────────────────────────┼────────────────────────┼───────────────┤\n",
       "│ conv2d_8 (\u001b[38;5;33mConv2D\u001b[0m)               │ (\u001b[38;5;45mNone\u001b[0m, \u001b[38;5;34m7\u001b[0m, \u001b[38;5;34m7\u001b[0m, \u001b[38;5;34m96\u001b[0m)       │        \u001b[38;5;34m41,568\u001b[0m │\n",
       "├─────────────────────────────────┼────────────────────────┼───────────────┤\n",
       "│ max_pooling2d_8 (\u001b[38;5;33mMaxPooling2D\u001b[0m)  │ (\u001b[38;5;45mNone\u001b[0m, \u001b[38;5;34m3\u001b[0m, \u001b[38;5;34m3\u001b[0m, \u001b[38;5;34m96\u001b[0m)       │             \u001b[38;5;34m0\u001b[0m │\n",
       "├─────────────────────────────────┼────────────────────────┼───────────────┤\n",
       "│ flatten_2 (\u001b[38;5;33mFlatten\u001b[0m)             │ (\u001b[38;5;45mNone\u001b[0m, \u001b[38;5;34m864\u001b[0m)            │             \u001b[38;5;34m0\u001b[0m │\n",
       "├─────────────────────────────────┼────────────────────────┼───────────────┤\n",
       "│ dense_6 (\u001b[38;5;33mDense\u001b[0m)                 │ (\u001b[38;5;45mNone\u001b[0m, \u001b[38;5;34m512\u001b[0m)            │       \u001b[38;5;34m442,880\u001b[0m │\n",
       "├─────────────────────────────────┼────────────────────────┼───────────────┤\n",
       "│ dropout_4 (\u001b[38;5;33mDropout\u001b[0m)             │ (\u001b[38;5;45mNone\u001b[0m, \u001b[38;5;34m512\u001b[0m)            │             \u001b[38;5;34m0\u001b[0m │\n",
       "├─────────────────────────────────┼────────────────────────┼───────────────┤\n",
       "│ dense_7 (\u001b[38;5;33mDense\u001b[0m)                 │ (\u001b[38;5;45mNone\u001b[0m, \u001b[38;5;34m256\u001b[0m)            │       \u001b[38;5;34m131,328\u001b[0m │\n",
       "├─────────────────────────────────┼────────────────────────┼───────────────┤\n",
       "│ dropout_5 (\u001b[38;5;33mDropout\u001b[0m)             │ (\u001b[38;5;45mNone\u001b[0m, \u001b[38;5;34m256\u001b[0m)            │             \u001b[38;5;34m0\u001b[0m │\n",
       "├─────────────────────────────────┼────────────────────────┼───────────────┤\n",
       "│ dense_8 (\u001b[38;5;33mDense\u001b[0m)                 │ (\u001b[38;5;45mNone\u001b[0m, \u001b[38;5;34m2\u001b[0m)              │           \u001b[38;5;34m514\u001b[0m │\n",
       "└─────────────────────────────────┴────────────────────────┴───────────────┘\n"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre style=\"white-space:pre;overflow-x:auto;line-height:normal;font-family:Menlo,'DejaVu Sans Mono',consolas,'Courier New',monospace\"><span style=\"font-weight: bold\"> Total params: </span><span style=\"color: #00af00; text-decoration-color: #00af00\">626,946</span> (2.39 MB)\n",
       "</pre>\n"
      ],
      "text/plain": [
       "\u001b[1m Total params: \u001b[0m\u001b[38;5;34m626,946\u001b[0m (2.39 MB)\n"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre style=\"white-space:pre;overflow-x:auto;line-height:normal;font-family:Menlo,'DejaVu Sans Mono',consolas,'Courier New',monospace\"><span style=\"font-weight: bold\"> Trainable params: </span><span style=\"color: #00af00; text-decoration-color: #00af00\">626,946</span> (2.39 MB)\n",
       "</pre>\n"
      ],
      "text/plain": [
       "\u001b[1m Trainable params: \u001b[0m\u001b[38;5;34m626,946\u001b[0m (2.39 MB)\n"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre style=\"white-space:pre;overflow-x:auto;line-height:normal;font-family:Menlo,'DejaVu Sans Mono',consolas,'Courier New',monospace\"><span style=\"font-weight: bold\"> Non-trainable params: </span><span style=\"color: #00af00; text-decoration-color: #00af00\">0</span> (0.00 B)\n",
       "</pre>\n"
      ],
      "text/plain": [
       "\u001b[1m Non-trainable params: \u001b[0m\u001b[38;5;34m0\u001b[0m (0.00 B)\n"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "model1 = tuner.hypermodel.build(best_hps)\n",
    "model1.summary()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 53,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/100\n",
      "\u001b[1m115/115\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 59ms/step - accuracy: 0.6039 - f1_score: 0.7542 - loss: 1.7127\n",
      "Epoch 1: val_f1_score improved from -inf to 0.92647, saving model to ./best_model/cnn1_weighted_f1.weights.h5\n",
      "\u001b[1m115/115\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m11s\u001b[0m 72ms/step - accuracy: 0.6050 - f1_score: 0.7544 - loss: 1.7090 - val_accuracy: 0.9274 - val_f1_score: 0.9265 - val_loss: 0.6664\n",
      "Epoch 2/100\n",
      "\u001b[1m110/115\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m━\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - accuracy: 0.8824 - f1_score: 0.8868 - loss: 0.6247\n",
      "Epoch 2: val_f1_score improved from 0.92647 to 0.93265, saving model to ./best_model/cnn1_weighted_f1.weights.h5\n",
      "\u001b[1m115/115\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m1s\u001b[0m 6ms/step - accuracy: 0.8831 - f1_score: 0.8874 - loss: 0.6216 - val_accuracy: 0.9338 - val_f1_score: 0.9326 - val_loss: 0.4240\n",
      "Epoch 3/100\n",
      "\u001b[1m111/115\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m━\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - accuracy: 0.9066 - f1_score: 0.9097 - loss: 0.4274\n",
      "Epoch 3: val_f1_score improved from 0.93265 to 0.93648, saving model to ./best_model/cnn1_weighted_f1.weights.h5\n",
      "\u001b[1m115/115\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m1s\u001b[0m 5ms/step - accuracy: 0.9067 - f1_score: 0.9098 - loss: 0.4263 - val_accuracy: 0.9363 - val_f1_score: 0.9365 - val_loss: 0.3389\n",
      "Epoch 4/100\n",
      "\u001b[1m112/115\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m━\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - accuracy: 0.9105 - f1_score: 0.9131 - loss: 0.3481\n",
      "Epoch 4: val_f1_score improved from 0.93648 to 0.94627, saving model to ./best_model/cnn1_weighted_f1.weights.h5\n",
      "\u001b[1m115/115\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m1s\u001b[0m 6ms/step - accuracy: 0.9104 - f1_score: 0.9129 - loss: 0.3480 - val_accuracy: 0.9465 - val_f1_score: 0.9463 - val_loss: 0.2626\n",
      "Epoch 5/100\n",
      "\u001b[1m110/115\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m━\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - accuracy: 0.9232 - f1_score: 0.9251 - loss: 0.2987\n",
      "Epoch 5: val_f1_score did not improve from 0.94627\n",
      "\u001b[1m115/115\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m1s\u001b[0m 5ms/step - accuracy: 0.9232 - f1_score: 0.9251 - loss: 0.2983 - val_accuracy: 0.9376 - val_f1_score: 0.9363 - val_loss: 0.2519\n",
      "Epoch 6/100\n",
      "\u001b[1m113/115\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m━\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - accuracy: 0.9191 - f1_score: 0.9214 - loss: 0.2598\n",
      "Epoch 6: val_f1_score did not improve from 0.94627\n",
      "\u001b[1m115/115\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m1s\u001b[0m 5ms/step - accuracy: 0.9191 - f1_score: 0.9214 - loss: 0.2599 - val_accuracy: 0.9121 - val_f1_score: 0.9086 - val_loss: 0.2756\n",
      "Epoch 7/100\n",
      "\u001b[1m111/115\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m━\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - accuracy: 0.9196 - f1_score: 0.9215 - loss: 0.2522\n",
      "Epoch 7: val_f1_score did not improve from 0.94627\n",
      "\u001b[1m115/115\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m1s\u001b[0m 5ms/step - accuracy: 0.9196 - f1_score: 0.9215 - loss: 0.2522 - val_accuracy: 0.9350 - val_f1_score: 0.9332 - val_loss: 0.2379\n",
      "Epoch 8/100\n",
      "\u001b[1m113/115\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m━\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - accuracy: 0.9363 - f1_score: 0.9377 - loss: 0.2306\n",
      "Epoch 8: val_f1_score did not improve from 0.94627\n",
      "\u001b[1m115/115\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m1s\u001b[0m 5ms/step - accuracy: 0.9361 - f1_score: 0.9375 - loss: 0.2306 - val_accuracy: 0.9439 - val_f1_score: 0.9434 - val_loss: 0.2051\n",
      "Epoch 9/100\n",
      "\u001b[1m111/115\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m━\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - accuracy: 0.9331 - f1_score: 0.9346 - loss: 0.2163\n",
      "Epoch 9: val_f1_score did not improve from 0.94627\n",
      "\u001b[1m115/115\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m1s\u001b[0m 5ms/step - accuracy: 0.9328 - f1_score: 0.9344 - loss: 0.2165 - val_accuracy: 0.9439 - val_f1_score: 0.9430 - val_loss: 0.2027\n",
      "Epoch 10/100\n",
      "\u001b[1m111/115\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m━\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - accuracy: 0.9471 - f1_score: 0.9481 - loss: 0.1900\n",
      "Epoch 10: val_f1_score did not improve from 0.94627\n",
      "\u001b[1m115/115\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m1s\u001b[0m 5ms/step - accuracy: 0.9467 - f1_score: 0.9477 - loss: 0.1905 - val_accuracy: 0.9299 - val_f1_score: 0.9278 - val_loss: 0.2280\n",
      "Epoch 11/100\n",
      "\u001b[1m113/115\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m━\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - accuracy: 0.9281 - f1_score: 0.9300 - loss: 0.2025\n",
      "Epoch 11: val_f1_score improved from 0.94627 to 0.94897, saving model to ./best_model/cnn1_weighted_f1.weights.h5\n",
      "\u001b[1m115/115\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m1s\u001b[0m 6ms/step - accuracy: 0.9281 - f1_score: 0.9300 - loss: 0.2025 - val_accuracy: 0.9490 - val_f1_score: 0.9490 - val_loss: 0.1712\n",
      "Epoch 12/100\n",
      "\u001b[1m110/115\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m━\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - accuracy: 0.9414 - f1_score: 0.9426 - loss: 0.1818\n",
      "Epoch 12: val_f1_score did not improve from 0.94897\n",
      "\u001b[1m115/115\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m1s\u001b[0m 5ms/step - accuracy: 0.9410 - f1_score: 0.9422 - loss: 0.1825 - val_accuracy: 0.9490 - val_f1_score: 0.9484 - val_loss: 0.1832\n",
      "Epoch 13/100\n",
      "\u001b[1m111/115\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m━\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - accuracy: 0.9359 - f1_score: 0.9373 - loss: 0.1960\n",
      "Epoch 13: val_f1_score did not improve from 0.94897\n",
      "\u001b[1m115/115\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m1s\u001b[0m 5ms/step - accuracy: 0.9359 - f1_score: 0.9372 - loss: 0.1962 - val_accuracy: 0.9490 - val_f1_score: 0.9488 - val_loss: 0.1707\n",
      "Epoch 14/100\n",
      "\u001b[1m111/115\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m━\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - accuracy: 0.9416 - f1_score: 0.9429 - loss: 0.1891\n",
      "Epoch 14: val_f1_score did not improve from 0.94897\n",
      "\u001b[1m115/115\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m1s\u001b[0m 6ms/step - accuracy: 0.9413 - f1_score: 0.9425 - loss: 0.1895 - val_accuracy: 0.9478 - val_f1_score: 0.9477 - val_loss: 0.1585\n",
      "Epoch 15/100\n",
      "\u001b[1m110/115\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m━\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - accuracy: 0.9394 - f1_score: 0.9406 - loss: 0.1775\n",
      "Epoch 15: val_f1_score did not improve from 0.94897\n",
      "\u001b[1m115/115\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m1s\u001b[0m 5ms/step - accuracy: 0.9389 - f1_score: 0.9402 - loss: 0.1781 - val_accuracy: 0.9490 - val_f1_score: 0.9484 - val_loss: 0.1829\n",
      "Epoch 16/100\n",
      "\u001b[1m111/115\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m━\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - accuracy: 0.9404 - f1_score: 0.9418 - loss: 0.1804\n",
      "Epoch 16: val_f1_score improved from 0.94897 to 0.95166, saving model to ./best_model/cnn1_weighted_f1.weights.h5\n",
      "\u001b[1m115/115\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m1s\u001b[0m 6ms/step - accuracy: 0.9402 - f1_score: 0.9416 - loss: 0.1809 - val_accuracy: 0.9516 - val_f1_score: 0.9517 - val_loss: 0.1560\n",
      "Epoch 17/100\n",
      "\u001b[1m111/115\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m━\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - accuracy: 0.9425 - f1_score: 0.9438 - loss: 0.1765\n",
      "Epoch 17: val_f1_score improved from 0.95166 to 0.95367, saving model to ./best_model/cnn1_weighted_f1.weights.h5\n",
      "\u001b[1m115/115\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m1s\u001b[0m 6ms/step - accuracy: 0.9421 - f1_score: 0.9434 - loss: 0.1771 - val_accuracy: 0.9541 - val_f1_score: 0.9537 - val_loss: 0.1667\n",
      "Epoch 18/100\n",
      "\u001b[1m111/115\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m━\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - accuracy: 0.9404 - f1_score: 0.9416 - loss: 0.1768\n",
      "Epoch 18: val_f1_score did not improve from 0.95367\n",
      "\u001b[1m115/115\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m1s\u001b[0m 5ms/step - accuracy: 0.9400 - f1_score: 0.9412 - loss: 0.1774 - val_accuracy: 0.9490 - val_f1_score: 0.9489 - val_loss: 0.1644\n",
      "Epoch 19/100\n",
      "\u001b[1m110/115\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m━\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - accuracy: 0.9416 - f1_score: 0.9429 - loss: 0.1915\n",
      "Epoch 19: val_f1_score did not improve from 0.95367\n",
      "\u001b[1m115/115\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m1s\u001b[0m 5ms/step - accuracy: 0.9412 - f1_score: 0.9425 - loss: 0.1918 - val_accuracy: 0.9478 - val_f1_score: 0.9479 - val_loss: 0.1542\n",
      "Epoch 20/100\n",
      "\u001b[1m111/115\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m━\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - accuracy: 0.9461 - f1_score: 0.9471 - loss: 0.1583\n",
      "Epoch 20: val_f1_score did not improve from 0.95367\n",
      "\u001b[1m115/115\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m1s\u001b[0m 5ms/step - accuracy: 0.9459 - f1_score: 0.9469 - loss: 0.1590 - val_accuracy: 0.9516 - val_f1_score: 0.9517 - val_loss: 0.1500\n",
      "Epoch 21/100\n",
      "\u001b[1m111/115\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m━\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - accuracy: 0.9410 - f1_score: 0.9423 - loss: 0.1700\n",
      "Epoch 21: val_f1_score did not improve from 0.95367\n",
      "\u001b[1m115/115\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m1s\u001b[0m 5ms/step - accuracy: 0.9407 - f1_score: 0.9420 - loss: 0.1706 - val_accuracy: 0.9529 - val_f1_score: 0.9524 - val_loss: 0.1719\n",
      "Epoch 22/100\n",
      "\u001b[1m105/115\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m━━\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - accuracy: 0.9325 - f1_score: 0.9337 - loss: 0.1914\n",
      "Epoch 22: val_f1_score did not improve from 0.95367\n",
      "\u001b[1m115/115\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m1s\u001b[0m 6ms/step - accuracy: 0.9323 - f1_score: 0.9335 - loss: 0.1920 - val_accuracy: 0.9414 - val_f1_score: 0.9421 - val_loss: 0.1649\n",
      "Epoch 23/100\n",
      "\u001b[1m110/115\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m━\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - accuracy: 0.9437 - f1_score: 0.9450 - loss: 0.1642\n",
      "Epoch 23: val_f1_score did not improve from 0.95367\n",
      "\u001b[1m115/115\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m1s\u001b[0m 5ms/step - accuracy: 0.9432 - f1_score: 0.9446 - loss: 0.1652 - val_accuracy: 0.9529 - val_f1_score: 0.9528 - val_loss: 0.1613\n",
      "Epoch 24/100\n",
      "\u001b[1m111/115\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m━\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - accuracy: 0.9431 - f1_score: 0.9444 - loss: 0.1690\n",
      "Epoch 24: val_f1_score improved from 0.95367 to 0.95519, saving model to ./best_model/cnn1_weighted_f1.weights.h5\n",
      "\u001b[1m115/115\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m1s\u001b[0m 6ms/step - accuracy: 0.9430 - f1_score: 0.9443 - loss: 0.1693 - val_accuracy: 0.9554 - val_f1_score: 0.9552 - val_loss: 0.1524\n",
      "Epoch 25/100\n",
      "\u001b[1m114/115\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m━\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - accuracy: 0.9424 - f1_score: 0.9436 - loss: 0.1672\n",
      "Epoch 25: val_f1_score did not improve from 0.95519\n",
      "\u001b[1m115/115\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m1s\u001b[0m 6ms/step - accuracy: 0.9424 - f1_score: 0.9435 - loss: 0.1675 - val_accuracy: 0.9516 - val_f1_score: 0.9517 - val_loss: 0.1501\n"
     ]
    }
   ],
   "source": [
    "model_file = \"./best_model/cnn1_weighted_f1.weights.h5\"\n",
    "checkpoint = tf.keras.callbacks.ModelCheckpoint(model_file, \"val_f1_score\", 1, True, True, \"max\")\n",
    "history_model1 = model1.fit(\n",
    "    X_train,\n",
    "    y_train_onehot,\n",
    "    epochs=100,\n",
    "    validation_data=(X_valid, y_valid_onehot),\n",
    "    callbacks=[early_stop, checkpoint],\n",
    "    class_weight=class_weights,\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 54,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 1200x800 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAABEwAAAKvCAYAAABwAVqsAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjkuMiwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy8hTgPZAAAACXBIWXMAAA9hAAAPYQGoP6dpAACuO0lEQVR4nOzdd3xUZdrG8d+ZPumEEnrvRQFBBAQUxVVQd+0NdBfsCqivKK6urO4quqwV29rWgmvH3hsdKYqFjiCd0EJ6pp7z/jFhQkgCCSSZlOv7+UAyZ86cuSf9XOd57sfIzMy0EBERERERERGRKFusCxARERERERERqWkUmIiIiIiIiIiIHESBiYiIiIiIiIjIQRSYiIiIiIiIiIgcRIGJiIiIiIiIiMhBFJiIiIiIiIiIiBxEgYmIiIiIiIiIyEEUmIiIiIiIiIiIHESBiYiIiIiIiIjIQRSYiIiIiIiIiIgcRIGJiIiIVLnrrruOlJSUEv82bNgAwPz587nooovo2rUrKSkpfPzxx4c9Zjgc5uGHH6Z///40bdqUtm3bcuqppzJjxoyqfjkiIiJSDzhiXYCIiIjUD6eeeipPPvlksW2NGjUCID8/n169enHZZZdx+eWXl+t4U6dO5eWXX2batGn07t2bnJwcli1bRmZmZmWXHhUIBHC5XFV2fBEREak5FJiIiIhItXC5XKSlpZV634gRIxgxYkSFjvf5558zbtw4/vSnP0W39erVq9g+pmny+OOP8/LLL7Nt2zaaNGnCn//8Z2699VYAVqxYweTJk1myZAler5ezzz6b++67j4SEBCAyMiYrK4t+/frx7LPP4nQ6+fXXX9m+fTt33nkn3377LTabjRNOOIEHHniANm3aVOg1iIiISM2lKTkiIiJSK6WlpTFnzhz27NlT5j733HMPjz76KJMmTWLRokU899xzNGnSBIiMajn//PNJSUnh22+/5aWXXmLWrFlMmjSp2DHmzJnDmjVreO+993jzzTfJz8/nrLPOIj4+nk8//ZTPP/+chIQEzj//fAKBQJW+ZhEREak+GmEiIiIi1eKLL76gRYsW0dunnnoqL7/88hEf77777uOKK66gc+fOdO3alQEDBjBy5MjoSJWcnByeeeYZpk2bxqWXXgpAu3btGDhwIABvv/02BQUFPPPMM8THxwMwbdo0Lr74Yu65555osBIXF8f06dOjU3FeffVVbDYb06dPxzAMAJ588knatGnDvHnzGD58+BG/JhEREak5FJiIiIhItRgyZAgPP/xw9HZcXNxRHa9r164sXLiQn376iYULF7JgwQIuvvhiLr30UqZPn87atWvx+/0MHTq01MevWbOGnj17RsMSgAEDBmCaJuvWrYsGJt27dy/Wt+Tnn39mw4YNtGzZstjxfD4fv//++1G9JhEREak5FJiIiIhItYiLi6N9+/aVekybzUbfvn3p27cvN9xwA2+++SbXXHMN//d//4fH4znkYy3Lio4QOdiB2w8OdkzTpHfv3jz33HMlHtewYcMjeBUiIiJSEykwERERkTqjS5cuQKQ/SYcOHfB6vcyZM4e2bduW2Ldr1668/vrr5OXlRUeZLFq0CJvNRseOHct8jmOPPZaZM2fSqFEjkpKSquR1iIiISOyp6auIiIjEXG5uLr/88gu//PILAJs2beKXX35hy5YtZT7m8ssv58knn2Tp0qVs3ryZuXPnMmnSJDp27Ejnzp3xeDxMnDiRu+++m9dff53ff/+dJUuW8MorrwBwwQUX4PF4uO6661i5ciVz5szhtttu46KLLopOxynNBRdcQMOGDbn00ktZsGABGzduZN68edx+++1s27atcj8wIiIiEjMaYSIiIiIxt2zZMs4666zo7TvvvBOASy65hKeffrrUx5xyyim88847PPLII2RnZ9OkSROGDh3K5MmTcTgif+LcdtttOBwO7r//ftLT00lLS2Ps2LFAZKrNu+++y+TJkxk+fHixZYUPJS4ujk8//ZQpU6YwZswYcnNzadasGcOGDSMxMbEyPhwiIiJSAxiZmZlWrIsQEREREREREalJNCVHREREREREROQgCkxERERERERERA6iwERERERERERE5CAKTEREREREREREDqLARERERERERETkIApMREREREREREQOosBEREREREREROQgCkxERERERERERA6iwERERERERERE5CAKTEREREREREREDqLARERERERERETkIApMREREREREREQOosBEREREREREROQgCkxERERERERERA7iiHUBVSkQCJCdnY1lWbEuRSqBYRgkJSXhcrliXYqIiIiIiIjUcXU2MAkEAmRlZdGwYUNsNg2kqQtM02Tv3r0kJycrNBEREREREZEqVWeThOzsbIUldYzNZqNhw4ZkZ2fHuhQRERERERGp4+psmmBZlsKSOshms2mKlYiIiIiIiFQ5JQoiIiIiIiIiIgdRYCIiIiIiIiIichAFJnXcqFGjmDx5cqzLEBEREREREalV6uwqObVNSkrKIe+/5JJLePrppyt83BkzZuBw6NMsIiIiIiIiUhE6kz6E9VkhZqzLY3NumNYJdkZ3iqdDctV8yNasWRN9f+bMmUydOpUlS5ZEt3k8nmL7B4NBnE7nYY/boEGDyitSREREREREpJ7QlJwyzFiXR//3dvL48lze21jA48tz6f/eTl5bl1clz5eWlhb9l5SUVGybz+ejTZs2vPfee4waNYq0tDTefPNNMjIyGDduHN27d6dZs2YMGjSId955p9hxD56S06tXLx566CFuuOEGWrZsSc+ePXnppZeq5DWJiIiIiIiI1FYKTEqxPivEhPmZmBaELYq9HT8/kw3ZoZjUNWXKFK655hoWLVrEKaecgs/no3fv3rzxxhssXLiQK664gmuuuYalS5ce8jhPPPEEffr0Yc6cOYwbN45bbrmFtWvXVtOrEBEREREREan56tWUnCeW5/DUitzD7pcdMDGt0u8zLRj6wU6SXIfPmq7vkcCNPRMrWmaZrrvuOs4+++xi28aPHx99/5prruGbb77h/fffp1+/fmUeZ8SIEVx55ZUA3HTTTTz11FPMmzePzp07V1qtIiIiIiIiIrVZvQpMcoIW2/PNoz5ObghyQ4c/Tk6wjNTlCPXp06fY7XA4zCOPPMLMmTPZsWMHgUAAv99PXFzcIY/Ts2fP6PuGYdCkSRN2795dqbWKiIiIiIiI1Gb1KjBJdBo0jzv8yJDsgEnuIWbdJDgo1wiTRKdRkfIO6+Ag5IknnuCpp55i6tSpdO/enfj4eO644w4CgcAhj3PwqjmGYWCaRx8kiYiIiIiIiNQV9SowubFnYrmmyKzPCtH/vZ2lTsuxGTDnj2m0T4r9h27BggWMHDmSiy66CADTNFm/fr2m1oiIiIiIiIgcJTV9LUWHZAfTB6dgM8BuUOzt9MEpNSIsAWjfvj2zZs1i0aJFrFmzhptuuoldu3bFuiwRERERERGRWq9mnPnXQJd1imdgmptX1+axOTdM6wQ7YzrH15iwBOC2225j06ZNnHfeeXi9Xq644gpGjhxJdnZ2rEsTERERERERqdWMzMzMyu1MWkPs3r2bxo0bx7oMqQL63IqIiIiIiEhVqznDJUREREREREQEI30rzjmfYuxJx2rUlODQkVhNW8a6rHpHgYmIiIiIiIhIDeGY8xnuF6eBAViAAc5P38A/bhKhIWfEurx6RU1fRURERERERGoAI30r7henYVgmhmlG32KZuF+YhrFza6xLrFcUmIiIiIiIiIjEWm4WrndfIDKspDij8D/n7E+ru6p6TVNyRERERERERKpTQT62TWuxb1iNbeMa7BvWYNu9/dCPMU0ci2cR7tiDcI++4PZWT631mAITERERERERkaoS8GPbsj4Sjvy+JvJvxyYMq2IL1hqAsXs73sfuxHK6CHfvS6j3QMK9B2KlNqma2us5LSsstY4+tyIiIlLTaEULEQEgFMK27Xdsv6/B/vsabL+vxrZ1A0Y4fMiHWS43ZutOhJs0x7nwK7CsyDSc/fcXvjVKezAQbtOJcO9BhHoPxGzbGWzqvlEZFJhIraPPrYiIiNQkpa1ogYVWtBCp60wTI31LYTCyBvvvq7FtWocRDBzyYZbdjtmyA2b7LoTbdcVs1wWzRVuwRyaAOOZ+hvuFaWAYYFlFP1OuuAmrQSMcyxZi/2khtsw9pZeV0pDwsScQ6j2IcI/jwO2p5BdefygwkVpHn1sRERGpEfJysP+6CM8z95UYWm8BGDbyH3wFK00jTURqPcvC2JNeFIxsWI1941oMX/6hH2YYmM3bREKRdl0jAUmr9uByH/Jxxs6tOGcfMGpt2MjiP0ssC9vGtTh+WoD9p4XYN64t/fk1deeoKDCpQ0aNGkWvXr144IEHAOjVqxfXXXcd119/fZmPSUlJYcaMGZx55plH9dyVdZzyqI+fWxEREYkhXz62bZsiw+y3bcS29ffIvzKu7u5nGQbBk84i8OdbqqlQESmvw02jMzL3HhSOrMHIyTrscc0mzaOjRsLtumK27QSeuKp8KZF6M3Zj//n7SICy4ocyR7lEp+70GYjZRlN3DkdNX2uIiy66CJ/PxwcffFDivsWLF3Paaacxa9YsevfuXe5jfvfdd8TFVe4359SpU/nkk0+YN29ese1r1qwhJSWlUp9LREREpFoF/Nh2bC4KRbYVBiN70o/ocIZl4fzuQ2x7dhA8+WzCvQdGh9yLVAb1zjkypU2jc376OqG+J2JYVqTvyL5DB6IAZoNGkVEj7btG3rbtDAlJVV5/aazUxoROPovQyWeB34d91Y+lTt2xb1qHfdM6XB+8rKk75aCf2IdQnT+AxowZw5gxY9i8eTOtW7cudt+MGTPo1atXhcISgEaNGlVihYeWlpZWbc8lIiIiclRCIYydW7Ft24h9fzCy7XeM9G0YllmuQ1jxSZgt20VGn2z+rczVLgzA8esSHL8uwUxpRGjYKILDRmE11LB4OTqln/S/od45h2Gkb8X94rTI9/r+b9vCt84f5pb5OCsxufjIkXZdsFIaVn3BR8LtIdx7EOHeg4pP3Vm2EPumoqk7tsy92GZ/gnP2J0VTd/oMInzsQKxUjegHBSZlqu4fQKeffjqNGzfmf//7H5MnT45uz8/P57333uOGG25g3LhxLFy4kH379tGuXTtuueUWzj///DKPefCUnPXr13PjjTfy448/0rZt2+jUnQNNmTKFjz/+mO3bt9OkSRMuuOACbr/9dpxOJ6+99hoPPvggQHQ0yZNPPslll11WYkrOihUrmDx5MkuWLMHr9XL22Wdz3333kZCQAMB1111HVlYWAwcO5IknniAQCHDeeecxdepUnE5npXxMRUREpO44ogtZpomxe0fRSJH9I0d2bMYIh8r1vJbHi9miHWaLtpgt20X+tWiHlZwKhoGRvpW4yZdjUdqKFgZWg4bRK9W2zD24PngZ54evEu49kODJZxHu1R9s9iP5kEg1iekoDtMEXz5GXg5Gfi5Gfi7k5WLbsQnXOy9gYBU76bcA9wvTCHfupd45ZXDO+bToHK8MljeecNvOhT1HukTCkUZNI01YaxvDwGzXhUC7LnDOX8qcumMEAzh+/h7Hz98Dh566U59GNikwKUVZqWNV/gByOBxcfPHF/O9//+P222/HKPxmfP/99wkEAlx++eW8++67TJw4kaSkJL744guuueYa2rZtS79+/Q57fNM0GTNmDKmpqXz11Vfk5ORwxx13lNgvISGBp556iqZNm7Jy5UomTpxIYmIiEydO5Nxzz2XVqlV88803vP/++wAkJZUccpafn8/5559Pv379+Pbbb9m9ezcTJkxg0qRJPP3009H95s2bR9OmTfnoo4/YsGEDY8eOpVevXlxxxRVH+FEUERGRuuiwF7IsCyNjd9E0mv0ByfZNGAF/uZ7Dcrowm7c9KBhpi9Uw7ZAnSVbTlvjHTcL9wjSsg1e0GDeJ0OA/YF++BOd3H2FftgDDMjEsE8ey+TiWzcdslEbwpLMIDTmj5l6trscq5SJqwB8JOgoDj/3hBwe8X3xbLkb+/tt55R71BIVlGuCc/SmBC68+kpdc5xm7tkWCqFJYhkH4mAH4brq/zvb2KDF1Z+WPOJYtwP7zQmyZe6P7lZy6M5BQn0EY+/bgfuXRejOyqV4FJs7P38L5+VuH3c8oyINSfjAZgGWZxN19FZY3/rDHCZ5+IcHTLyx3faNHj+bxxx9n7ty5DB06FIhMxznrrLNo3rw548ePj+57zTXXRIOL8gQms2bNYs2aNfzyyy+0aNECgLvvvrvECJVJkyZF32/Tpg3r1q1j5syZTJw4Ea/XS3x8PHa7/ZBTcN5++20KCgp45plniI+PfJymTZvGxRdfzD333EOTJpEhqMnJyUybNg273U7nzp057bTTmD17tgITERERiTrkhaznH8T55bvYdu+I/P1WDpbdjtmsdSQYaXHAiJEmzY54pEdoyBmEO/cqc0WL8DEDCB8zACNjF47Zn+Kc/XHRqJM9O3G/8zyu9/5LuO+JkV4n3frU2ZO12uTQF1H/hZG5FxxOjLycojDkwPBjfxhymCVmK13hai5SiuxM7Gt/pcwI1DAwW7avP99/bg/hPoMI9xkEpolt07pDTN35GOfsj6Mrp9eXkU31KjAxCvLK1bznkMcA8BVg+ArK9XwV0blzZwYMGMCMGTMYOnQov//+OwsXLuS9994jHA7zyCOPMHPmTHbs2EEgEMDv95e7qeuaNWto2bJlNCwB6N+/f4n9PvjgA55++mk2bNhAXl4eoVCIxMTECr2ONWvW0LNnz2hYAjBgwABM02TdunXRwKRbt27Y7UV/mKSlpbFy5coKPZeIiIjUbc5ZH5W6ff8Jj33zb6Xebxk2rLQWRSNG9k+radoSHJU//ddKa3nYK/pWahOC5/yZ4Nmjsf+8COd3H2L/dTGGZWGEwziWzMaxZDZmWkuCJ51JcMjpkJhS6bVK+Ti/eJvS5m1EThYt3O88XyXPa9lsEJ+IFZeAFVf4Nj4R4hKw4hOw4hKwr/4J+/KlpffOsSyMgK9KaqvNjN078E6bhC0rA6DoxL+QVfhfcNjIGFRXA9hs5Zu6U8pD6/LIpnoVmFjeeMwGh2+EahTkRUKR0o4B4PGWa4RJefY52OjRo7ntttvIzs7mtddeo1WrVgwbNozHH3+cp556iqlTp9K9e3fi4+O54447CATKl1hbpfwwNQ4aXrpkyRLGjh3LHXfcwSmnnEJSUhLvvvsuTzzxRIVeg2VZJY5d2nM6HI4S95llDI8TERGResKyMHZsxvHLYuy/Lsa+ooyTwgOYjZoWD0ZatsNs1hpc7moquoLsDsJ9BxPuOxhj9w6csz/BMecTbFn7ALDt3Ir7zWdwvfsCof7DCJ58NmbnXrWzf0ItY2zfhOOHuTiWzsG+ce3hH1AGy+ONBh4UBh5WYeBR7HZhIMKB97u9h/1ch/oNK7V3DhQ2Gl62AOcn/yM48hJ93QC2TevwPHR7NCwx4xIwCvKwDFuJaXR1bYTEkSpt6o77tScwdm8vfYSORZ0c2VSvApPyTpGJNu+yzJKpo2Ej/97nquwb6ZxzzuGOO+7gnXfe4fXXX+fyyy/HMAwWLFjAyJEjueiii4BIT5L169fTuXPnch23a9eubN26lR07dtCsWTMgEpAc6Pvvv6dVq1bceuut0W1btmwpto/L5TpsqNG1a1def/118vLyoqNMFi1ahM1mo2PHjuWqV0REROqR/NzIPPpfFmNfvgTb3p3lephl2Aiedh6BS2+o4gKrjtW4GYHzryTwpyuw/zgf53cf4lj5IwBGKIhz4dc4F35NuHlbQiefRXDwaRBfsdG/cgiWhW3zbziWzsGxdA627ZvK9zDDINyzH6EhI6MjQKKBhzcBHFV7mlVm7xyz6PzF/daz2HbvwD9mYr1eztq+ahmex+6Kjv43m7WmYNI0CAXLnEYnBymcuhNatxznZ2+U3gPGINIYt46p8HfO+vXrue6669i7dy/Jyck89dRTdO3atdg+pmly991388033xAKhRgwYAAPP/wwLpeLTZs20bdvX7p16xbd/9VXX6Vdu3YALF26lJtuuomCggJatGjBs88+S9Om1fuBP1zzrqr8RkpISOCcc87h3nvvJTs7m0svvRSA9u3b89FHH7Fo0SJSUlJ48skn2bVrV7kDk5NOOolOnTpx7bXX8s9//pOcnBz+8Y9/FNunffv2bN26lXfffZe+ffvyxRdf8PHHHxfbp3Xr1mzatCnaCyUhIQG3u/jVmwsuuICpU6dy3XXXMXnyZPbs2cNtt93GRRddFJ2OIyIiIvVY4Vx5+6+Lcfy6BNtvyzHKuCBjJqdiFF4VLrkKDQRP+WPV1lpdHE7Cx59E+PiTMNK34PzuI5zzPsfIzQbAvn0j9tem43r7WUIDhhM8+SzM9t00euBImCa29SuLQpIyroqHm7ctDFBKXwHJP2ZiTE+wS+2dM3QkjiWzotOFnN99hJGxG9/1d4OnfFP56xL74ll4/nMfRigIQLhDdwpumQoJyQB1bvpIVQsOHYnz0zfq1XSmCgcmN910E1dccQWXXXYZH3zwAePHj+err74qts+rr77KihUrmD17Nk6nk/Hjx/PMM88wYcIEINLsc968eSWObVkWV111FY8//jhDhgxh+vTp3HnnnbzwwgtH+PKO3OGad1WlMWPG8OqrrzJ8+HBatWoFwG233camTZs477zz8Hq9XHHFFYwcOZLs7OxyHdNmszFjxgzGjx/PKaecQuvWrXnwwQc577zzovuMGjWK66+/nkmTJhEIBDjttNOYNGlSseWHzz77bD766CPOOusssrKyossKHyguLo53332XyZMnM3z48GLLCouIiEg9lZ2JY/mSyDSbX5dgy8ksdTfL6STcpTfhXscTOuZ4rGatccz7PCYXsmLFatqKwCXXEzhvHI6lcyIr7Kz9BQAj4Mc59zOccz8j3LojweFnEzrhVPDWv5PhCgmFsK/5CcfSudh/mBudmnEgyzAwO/UkdNxQQsediNW4GY65n9Xor73SeucEzxqN1agp7ucewAiHcPz8Pd77J+K7eSpWOdoT1BXOr2biem16dEpfqPdAfNdPAbcnxpXVXrEcWBArRmZm5qEnhR5g9+7dHHfccWzYsAGHw4FlWXTp0oWvvvqKNm3aRPebNGkSLVu2ZOLEiUCkkeiDDz7IggUL2LRpEyeffDIbNmwocfwff/yR66+/nu+/j6z9nJOTQ6dOndiyZQtOZ8Wac+3evZvGjRtX6DFSO+hzKyIiUguEQ5Gr+IW9SGyb1pXZi8Rs1opQr+MJ9zyecNdjSz2hMXZurdfD521bf8cx6yOc87/AyC++sIDl8RIaeGqk10mbTjGqsAYK+LGv+CEykmTZAoy8khcaLbudcNc+hPoPJdxncKlLO9fWrz37qmV4Hv9bZMUewExtgu//HsRs2S7GlVUxy8L17gu4PpoR3RQcOhL/n2+p11OTKlNt/Z44EhX6itm2bRvNmjWLNus0DIOWLVuydevWYoFJnz59ePnllxk7dixut5uZM2cW64WRk5PDySefTDgcZtSoUdx6663Y7Xa2bNkSHVEBkJiYSEJCAunp6cW2H8znK9kF2jRNNRCto0zTLPVzLiJSl9h2bsOz4Avse3cSbpiGb9AfMNNaHP6BIjFky9iFc8UPuFb+gHPVMmxlrBhour0Eu/Ym2KMfgR7HYR44790CSvs9n9yIgrMvL76tPv090KgZnH81nH057iVz8Mz5BOfGNQAYvoLIFJ7vPiLYriu+ISPx9x8Krvp3Jd3w5eP8dQnuZfNxLl+CzV9yZUvL6SLQ/TgCfQYTOGZAZAWa/erS1167bvhue5ik6Xdh37sLW8YuPP+8kZxr/0awa+9YV1c1wmESXnsM1/wvo5vyz7iY/D9eAcFQ5J8cvdr6PVHI4yn/z8YKR2wHr35S2uorl1xyCVu2bGHkyJHEx8czbNgw5syZA0DTpk1ZuXIljRs3Zt++ffzlL3/hiSeeiI5GKc/xD7Z9+3bC4XCxbS6Xq9wryEjt4vP5yj0VSUSkNkr9aT6tP3mZ6DhXDLxfvM3mM68g49jBMa6u5nNn7CT1p/m4M/fgT2lERu/B+FPTYl1WnWSEgiRsXkfS+uUkrl+Od8+OMvfNT2tFdoee5HToQV7LDlj7r/QWBOGgJvNyGK27wehueNM30+jH2TT4dRH2oB8A5++rcf6+Gu+bT5NxzED29h2Kr3GLOv19Yc/PJXndz6SsXkbihhXYwiVPisMuN9kdjyGza1+yO/bE3B8mZWRG/tVZBo4xt9HhzenE7diErSCPxMfuZPOZV7DvmIGxLq5SGUE/7WY+i2ddZPqahcHWP1zMnuOGw9atMa5Oagq73U779u3LvX+VTMk52Lvvvsvzzz/PZ599VuK+d955h7fffps333yz1Ck5HTt2ZOvWrYecklPaaIOsrCxN26ijdu/eTXJycqzLEBGpErad22gw5coSUxciK7UZ7Lv3BcwmzWNS2341efSLe/6XJLz6SKQRpmVF3+ZefjP+QafFurwa77CfW8vCtmsbrhVLca34AeeaXzAKT9QPZsYnEejeNzKKpHtfrOTUanoV9Y9RkId78Sw8cz7BsbXktPdQkxbYd2+vU98XRtZe3D8txPXjPJxrfym1abAZn0jg2IH4+w4m2LUPOF0xqLSG8BWQ9PwDuH5dFN2Ud/YYCkZeWicaBxu52SQ9OQXnhlUAWA4nOX+ZRKDf0BhXJjVRlY0wady4Mb169eLNN9/ksssu48MPP6R169YlwhKfz4fP5yMlJYW9e/fyyCOPcOeddwKRk92UlBScTid+v5+PPvqIY445BoDevXvj8/mYO3cuQ4YM4aWXXmLUqFGH7V9S2gvOycnBZrNV5OVJLWGz2Sr0RS4iUpu4Fn1TdFJzAIPIqMuUB2/GatwMy+MFtwfL7QWPF8vtAbe3cLs3st3twfIUfz9yX2TfI1n20jHnM9wvTjtw8AveL97BP24SoSFnVMrHoEyWBcEAhEMQDGKEAhAMQiiIEQpi7NyG59VHImHT/o+fZWEBCa88iq1H3zo7x7oylPm5HTMRq0HDoiV/d5c+isQybJgdukV6kfQ6HrNdZ7DZMQB3qY+QSuPxwB/Ow3faudg2rML57Yc4Fn2LEYyMtnbs2hbZ7+Dvi5cfwZWfg9myHVZSKlZSClZSA3DVzM+YsXtH4co2c7GtX1FqTxwzpSGh44YQ7jeUcJdjwO7ADtirv9yaxeMhcPN9GDOm4/z2AwDiP3wVV+Ze/FfcUuXLIFclY+9OvP+ehG3HZgAsTxy+m+7D1q0POmOQo1WhESYA69at4/rrrycjI4PExESefvppunXrxvjx4znjjDMYOXIku3btYtSoUdjtdsLhMNdddx1jx44F4MMPP2Tq1KnYbDbC4TBDhgzhn//8Z3Rp2sWLF3PzzTfj8/lo1qwZzz77LM2bV/xKmhqD1l363IpIXeZ+6l4ci78rszlmZbIczgNClkj4Yrk9hQHMwSGLN7I6x0evljn6JXDhNVhxCZHlG0PBYmFG5Hag8P0QhAIYhfdH94nuHyjlsUGMUobZl/u1AnjiMJu1xkpOLfzXACs5FfOg2/Vx6U0jfStxky/HsIpfpd//mS7r+rOZ0ojwMcdHQpIex8GBvSAktvJycM7/EtcHL0eXJi4vy+PFSmpQ+C8FK7HwbXJRqGIlpmAmN4D4JDjKi5RG+lacc4ovjWs1jYSbxraNkZDkh7nYN60r9fFm42aE+g0l1G9oZKllXTQtm2Xh/OxN3G8+E90U6tEP3/h7wBsfw8KOjG3rBjzTbsOWuQeILEPu+78H1fxYKk2FA5PaYs+ePaSmpmqUSR1jmiYZGRk0alR/lkQTkfrF9dazOD99vdTAxIJIh38zXC2BSn1luT2FJ4apWCmFgUpSgwOClsJw5QiuxB/qxLDSmGEoyMfIz42sjlGQh5Gfh1GQd8Dt3MhKKwV5GAW52LZswMjcW2Ywsp/lcBLu3Ivw/lEkLdvVieH8dVlVh7CWYcNKSi4eqiSmFIUtyQ2K7ktqUGIFpNJGNmFZhI89AdvObdFRAwcLt2hLeH9I0qqDvg4ryLHoO9zP3R8JroFwqw74bpmKldokxpWVn23NL3gf/WvRKkBpLSmYNA2rcbMYVyZ1SZ0NTAKBAFlZWTRs2FChSR1hmiZ79+4lOTkZl6sez0EVkTrN+G0lcf+4vsSJa2QUh438B1/BatICAn7w+zD8BRj+AvAVvvX7MHwF4C/A8PsKt/sit6P7FBTu4yv5mBgGMZZhgMMJTieWw1X0vj3ytvh9DnA4sZyF+zmc2DauwbZxbdlhk8sDQX+lvkYrLiEaoJgHBipJhW9TUqMnjo75X5VyYkjx6UymWRhi5BUFGoXBh1GQd0D4kQv5eUXbC7cZBXmRz2UlsgAzrQWBS28g3LV3vRyBU5u53noW52dvlNrjwzIMwj36Ee7UEyN7H0Z2JrbsfdH3S1uG92hZbk/R9B+XB/uqHw8b1O0XbtclMpLkuCFYzVpXem31jW3tL3gfvSv6eTYbNMJ3ywOYrTvGuLLDs/8wF8/T9xYFPu26UnDLA5CUEtvCpM6ps4EJREKT7Ozscq20IzWfYRgkJSUpLBGROs31yqO4vnkfiHT4xzBKP7GuCvt7hBQGK0VBTCRwcc7+FPvPC0sPJAyDcI/jCPU/KdJY0eHEioYcrsiKKE4nOFxF2x3OyLSg/bftRzeHfv+0Eiyz2AlYsbCpUVOMnCyMrIyif9n7MDKLbtuy90Xez8s5qnoOVNbUlv3breRUDL8Pw5dfac9ZUfsznBLbbTaCZ1xM4MKrq7skqQTl+r4oq7dPKISRk1kUoBwYpmTvi9yXlVF0OxSs1NotwOxyLKF+QyIhScO6sapPTWLs2Iz3ocnYdm8HCvt/3HgP4V79Y1xZ2RzffYj75UejUwhDvfrju/EehblSJep0YCIiIlKbGDs2E/fXP2OYJpbLRXDoKIycrMjUjWEjY96w9KhOvKqJY+5nuF+YdsBqIBx52BQMRE4Cs/YVD1iyMrAdHLhU8qiOirJcHqy4ePDGR0a9xMVjeRMKb8djeeMhLiFyX+G26L7eeIzsDOL+OrZGf27lyFXq90VZLAt8+ZHvl2iockCYUsHRK5ZhEO4zGN/Ef1ZOfVImI3sfnkf/in194QozNhv+v9xKaOjIGFd2EMvC+f7LuN9/KbopOOg0/ONuq9VNa6VmU2AiUkmqZV66iNRpnkfvxLFsPgD+c8cS/OPlMa6opGo58TpKxs6tOGcf8PO4OsKm/SeK+wOU/aFKZuS2be2vGHnZpY/ggMgonEZpWN6EwnAj/oBw4zDBhye+Uk4WasPnVo5cTL4vDiUUwvW/J3F++0GJZsOgkU3Vzu/D88w/cfw4L7op8MfLCZzzl5rRH8YM437lUZzffRTdFBh5MYELrlaTX6lSCkxEKkHpDcv0R6aIlJ991TK8D9wMRFYeyf/Xq5Glf2ugGnfiVQscso9EDTox1OdWqlNtGLVWr5hhXP97CtdX70Y3RUZwTIpMm4yVgB/P0/8oFub4L7me4OkXxq4mqTcUmIgcpUMuxahf9iJSHqaJ9+/XYt+0FgDflbcrbK1jdGIoUjqNbKp5nF+8jev1p6L9qkLd+uAbf29slg3Py8H76J3Y1/4CgGV34L9qMqGBp1Z/LVIvafySyFFyzvm01C55RuF/ztmfVndJIlLLOBZ+HQ1Lwq07Ehp8WowrkspmNW0ZuUpr2LBs9shSrDYbGDb84yYpLJF6KzTkDPIffIXgGRcROv4kgmdcTP6DrygsiaHgHy7Ad+M9kVXIAMeqZXjvG4+xJ71a6zAyduG9f0JRWOLx4rvlAYUlUq00wkTkKLmfuhfH4u/KWDXCRuj4k/Bff3cMKhORWiHgJ+720dgydgNQcNtDhHscF+OipKpoyouI1Ba231bgffSvGDlZAJjJqZFlh9t2rvLnNrZvwjttEraMXZHnTkzB938PYrbrUuXPLXIgtRMWORqmiW3vzlLDkggLq1HTai1JRGoX5xdvR8OS0LEnKCyp46y0ljWiV4mIyOGYHXuQ/7cnI8sO79yKLSsD7/0T8N3wd8LHnlBlz2v7bQXeh++IrqRkNm5OwaR/KVyWmNCUHJEj5S/A8+Tfsf+2osxdDMuCvGwopcmfiIiRlYHr49eAwmUcL7o2xhWJiIgUsdJakv+3Jwh37AmA4ffheeSvOL77sEqez/7TArwP3hINS8JtOlFw13SFJRIzCkxEjoCxdxfe+ybgWDoHAAsj8s9mi8xLP6CpiWvWx3ie/Dv4fTGqVkRqKtd7L2H4CgAIDTsTq0Xb2BYkIiJysMQUCm5/iFD/YQAYlonnpYdxvfVspV4UdMz5FM9jd2EE/ACEuvel4I5HsVIaVtpziFSUepiIVJBt/Uo8j92JLWsfAJYnDt/1d2M2bVlsXrrlcOD6cEZ09Zxwuy74brpfP/RFBABj20bi7hqLYZpYHi/5/3oNKzk11mWJiIiUzjRxvfUfXJ+9Gd0UHDAc/1WTobBB7BGxLJwfv4b7necPOO7J+K+64+iOK1IJFJiIVIBjwVe4X/wXRjAIFM6pvPn+Mq8K23/+Hs9T92L48iP7pzbBd/P9mK07VlfJIlJDeR6ejOPn7wHwn38lwbNGx7giERGRw3N+/R6uGdOLLgp2PoaCif+EhKSKH8wM43rtCVxfvxfdFDjtPAKX3AA2TYaQ2FNgIlIepolr5ou4PpoR3RTueiwFN94DiSmHfKhtywY8j9yBbe9OoHBJtOvuJtx7YFVWLPWEkb4V55wDVtwYOhKrqeb51nT2FT/g/df/AWCmNib/gVfB7YlxVSIiIuVj/3E+nqfvjU6fMZu1ouCWB7GaNC//QYIB3P+5H+eSWdFN/guvJjjyEjCMsh8nUo0UmIgcji8fz7NTcfwwN7opOOxM/JdPBIezXIcwMvfieewu7BtWAZHlhgOX3kBwxLn6hSBHzDHnM9wvTgMDsIi+9Y+bRGjIGTGuLkKBTinMMN4p12Df/BsAvqv/SmjwaTEuSkREpGJsG1ZHLgpmR6apm0kN8N08FbN918M/OD8Xz2N34Vj9E1DY+HzcbYROPL0KKxapOAUmIodg7N2J59G/Yt+8HjjKoCPgx/3s1GIpeuCUPxG47Eawa4VvqRgjfStxky+PDofdzwIwDArunI7ZvE1k7q/TFZNgrjYEOrHgmPc5nuceACDcpjMFf39Gw45FRKRWMnbvwPvQ7dh2bAbAcrnxXfc3wn1PLPsxmXvxPHRb0d/XLg++G6t2qWKRI6XARKQMtt9W4Hnsrmhqbnnj8V1/N+FjBhz5QUuZ2hPq1R/f9VMgLuFoS5Z6xPXWszg/ewOjnN3pLacTnC6s/QGK44D3D9huOV2RkVNOF5ar5P04nMX3LfZ+0XMYmXvxTr0psrT2gXUAGDbyH3ylfi4R6PcRd/tobPv2AFAw+RHC3frEuCgREZGjkJuN9/G/YV/zMwCWYRC4bHzkAuNBjPQteKdNwrYnPbJvQhIFtzyA2aF7tZYsUl4KTERK4Zj/Je4Xp2GECpu7NmlOwc1TsZq3qZzjz/sc94v/xgiHAAi3aIvv5qlYjZtVyvGl7nM/dS+Oxd+VCCRqA8tmI3jGxQQuvDrWpVQ75wev4J75IgCh3oPw3Xx/jCsSERGpBMEA7ucfxPn9N9FNgdMvJDjsTJzzPsfYkw4OJ46f5mPk5QJgNkqj4NZpWM1ax6pqkcPSPACRA5lhXO88j+uT16ObQt364Lvx75CQXGlPEzrxdMxGTfE+fjdGXjb2bRvx3ns9von/xOzYo9KeR+oo08S2e0eZYYkFWGktMJs0xwgGoPCfEQwe8P4Bb6ubReQPp3rGyNyL65P/AYVztS+6JsYViYiIVBKnC/81d2I1aorr49cAcH3+Fs7P34pMO7UssCz2TxAOt2yP79Z/YTVoFLuaRcpBgYnIfgX5eP5zH45l86ObgiefjX/0BHBU/reK2bU3+Xc/hfeRO7Clb8GWvQ/vAzfhv+oOQgOGV/rzSR1R2AtnfwPhg+2f8lLwfw+Wb8qLZUEoEqQYoQMClUCgRLhCKFg8aDnE+/bf12Ckb6HUzikGWI2aHsUHoXZyvfcSht8HRH62VNaINRERkRrBZiNwwVWYjZrifvlhjP0BiVmy35rvytsVlhzG+qwQM9blsTk3TOsEO6M7xdMhuWacvtfk2iqbpuSIELna7Xnkr9i3bgAiV38Dl95I8NRzqr5ZZm42nul3R7uEA/jPHUvw7DFaQUeKy87E+9id2H9bAUD0Os3+Kzc1qKnq/qa0WGax0GR/U9r8B1+tVz1MbFs34L3rSgzLxPLEkTftf5CUEuuyRESkHOrTyWFlcT/xdxxLZpV64aQ+T80trxnr8pgwP/PgvvlMH5zCZZ3iVVs10ne61Hu2tb/iefxv2HIyAbDi4vHdcA/hnv2qp4CEJHyTpuF+6WGccz8DwD3zRWzpW/GPvTXSVFPqPWPHZrwPTca2ezsAltuD7/opmM1a4Zx9wLK9w0bWiCDCatoS/7hJuF+YhmUAZiQ4MQCzQSOslPp1Vcn15n+iKxoFzrpMYYmISC1R2snhY8tz6+zJYaWx2SIX/kqbPlxPp+aW1/qsEBPmZ2KW8qG7YV4mn2/2keS2YQA2gwPeGhgG2AAK3xrR+43I32EHbN//2P2PM8p4THQ/A/b5wjy2PI/SRlyMn5/JwDQ37ZPqVsRQt16NSAU55n6G+78PRZuvmmktKbj5/upvPuVw4h93G2azVrjfehYA54Ivse1Jp2DCvZCYUr31SI1iW/0z3sfvwsjLAcBMaYTvlqmYbToB1NgrNKEhZxDu3CsS6OzchmPVMoy8bGwZu3G/9DD+q++oF6Oo7MuX4vhlEQBmwzSCp50f44pERKQ8DnXiWldPDiuL1ahp2YFJPZ2aWx5h02LK0qxSv+b2+2izr/oKqgADeHVtHlP6VV7fx5rAFusCRGLCDON642k8zz8YDUtC3fuSf/dTsevUbRgER11KwY33YLncANjX/kLcvddjbN8Um5ok5hwLvsI77dZoWBJu1YGCKU9Fw5KazkprSeDCq/GPv4f8u6ZjuT1AJBB0fPdhjKurBoU/a/YLnH8lFH5/i4hIzTZ9eU6p5/tQdHIopQsOHQkWJUYiWIX/BYeNjEFVNVem32T6rzn0fncnH9fQQORwLGBzbjjWZVQ6RaJS/xTk4Xnmnzh+WhjdFDjlTwQuvbFKmrtWVLj/MAoapuF59K/YsjKw7dpO3D9uwDf+XsLd+8a6PKkuloXzw1ejS9AChHodj++GKeCtnUOAreZt8I+7Hc9T9wDgfu0JzDadMTt0i3FlVccx7wvsW9YDEG7XhdAJp8S4IhEROZwluwI8vjyHjzaVfeIatuCjTQXqZ1KG4lNzjRK91mrC9OGaYE1mkP+szOON9fnkhw7fWtRuwOhOcVzfIwGzMJDa/9ayrOjIlOg2wLQsLAtMChcrOugxVuH2/fcXPTZyvAPve3N9Pp9t9mGWUpsBtE6wH/0HpYZR01epV4xd2/E8+lfs2zYChUt7jp5A6JQ/xbSu0hh7d+F55I7oyZZlt+O/4hZCw0bFuDKpcqEg7v8+hHPe59FNwZPPwj9mIthr/x9lrteewPXlOwCYqU3Iv/fZujntzF9A3G2jsWXuBSD/jscwux4b46JERKQ0pmXx+RYf05fnsnBnoEKPHd7czbiu8ZzeyoPdVvenmlaEsXNrjey1FkumZfH1Vj/PrMzl2+3+EvcPTnOyYGew1D4hNgOWnpsWs6lg67NC9H9vZ6lThmJdW1VRYCL1hm3NL3in/w0jJwsAKz4R3w1/J9zjuBhXdggF+XievhfHz99HNwVGXkzggqsjzbSk7snLwfPEFBwrf4xu8l90LcEzLqo7/T5CIbwP3IR93fLIzZ798f3fA2CrW1clnO+9hPv9lwAI9T0R38R/xrYgEREpwR+2eHN9Pk8sz2VtVqjYfY3cBnv9VqknrqVpGW9nbNd4xnSKo7G3bv1Ok6OXEzT537p8nl2Vy/rs4lNX4h0Gl3aK45pu8XRMdvLaujzG19CVaGpybVVBgYnUC445n+J+6eGi5q7NWlFw01SsprUg4TbDuF5/CteX70Y3hY4bgu+av4LbG8PCpLIZu3fgefgO7Ns3AmA5nfiuvpPw8SfFtK6qYOzbg/fuq7Bl7wMg8MfLCZw7NsZVVR5j3x7ibhuNEfBh2e3k3/8SVtNWsS5LREQKZfpNXlyTx39W5rKzoPgEgy7JDsb3SuCC9nG8syG/1JPDB45Pwh+GF9bksTGn+MmvywZ/auvlym7x9G/swqgrFzzkiPyeHeLZVbm8ti6f7GDxU+82CXau7p7A6E5xJLuKXwzdkB3i1bVFy1mP6RxfY0Zv1OTaKpsCE6nbzDCuN/+D6/O3optCPfpF+kDEJ8awsIpzfPM+7hmPY5iRX+rhtp3x3XQ/VoP6tTxrXWXbsBrPo3dgy4oECFZiMgU33Y/ZsUeMK6s69lXL8Pzr/6Jf0wU3TyXce2CMq6oc7hf+hXPOpwAERpxLYPSEGFckIiIAW3JDPLUil1fW5pN3UM+IQWkuJvRK4LSWHmwHhByHOjncP73ihdW5fLnVX2I0yjGpTq7sFs/57b3EOTQ6uL6wLIs5O/w8vTKPL7b4SnxdDG3m5tru8fyhpaZx1XQKTKTuys/F8/Q/ost5QuGJyyXX19o+EPZfFuF58h4MXz4AZmpjfDfdX2tWTJHS2X+Yi+eZf2IEIvNYzaatKLjlAay0FjGurOo5P30D95vPAJFpcvn3PIvVuFm1PPf6rBAz1hX9AVxZjftsm9fjvftKDMvCiosn71+v1c0eLSIitcgvewNMX57LzN8LCB9w9mMz4Kw2Hsb3TKRfY9dRPcfGnBAvrs7j1XV57PMXP8VKdhlc2jGOK7smqElsHZYfMnl7fQH/WZnLysziU7w8driwQxzXdEugR6ozRhVKRSkwkTrJ2LkN76N/xVa4HK9lt+MfM5HQyWfHuLKjZ9u6Ac8jd2DbsxMAy+3Bd93dhPsMinFlUmGWhfPLd3C9/hRG4bqF4c7HUDDxn5CQFOPiqoll4Zl+N44f5gIQbtOJgrueqPKld2esy2NCFc2/9UybhGP5EqCw/8zIi4+2XBEROQKWZTFru5/Hl+fy3UHNNT12uKxTPDf0SKj0qQQFIYv3fs/n+dV5/LgnWOJ+NYmte7bmhnh+dR4vry0ZlrWIs3Nlt3iu6BxHqke9bWobBSZS59hXLcMzfQpGXjZQ2Nx1/L2Eu/WJcWWVx8jKwPPYXdjXrwTAMgwCl1xP8LTz605j0LrODEdWi/n6veim4MBT8Y+7DZxHd4Wr1snPJe7v12LbuRWA4NCRkY9DFanKDu/2Xxbhfeh2AMxGTcmf+nKVhz8iIlJc0LR47/cCpi/P5deM4oFFqtvGVd3iuapbPI2q4eT1x90Bnl+dx7u/5+Mv3upETWJrOcuyWLQrwDMr8/hoU/GRSwADmri4tns8Z7bx4lQwVmspMJE6xTHrY9yvPIIRjvxGMpu1puDm++vm8mUBP+7nH8C56LvopuDJZ+MfPQEcGupZo/nyI9PFfloY3RT44+UEzvlLvQ28bFs24L33uui0JN/YSVW2hPY9S7N4fHluiT9sAOwGTOiZwJR+yRU/cDiE929XRpct9133N0InnHJ0xYqISLnlBE1eWZvP0yty2ZpXPJ1om2jnhh4JXNYpLia9RDJ8YWasy+eF1XlsylWT2NrMH7aY+Xtk2s1Pe4sHck4bnNvOy7XdE+jTqJ5dAKujFJhI3RAO4Xrj6eIryfQ6Ht/1d0NcQgwLq2Kmieu9/+L68NXoplDP/pGmtnX5dddixr49eB75K/ZNa4HC6WJ/uZXQkDNiXFnsORZ8hec/9wGRFYIK7noSs23nSn+ecbMymPl7QanLRNoMOKetlxdOSq3wcR2zPsbz338DEO7QjYK/PVVvAzARkeq0Mz/Mf1bl8sLqPLICxX+6923kZELPRM5qUzOmv+xvEvv86ly+UpPYWmVnfpgX1+Tx3zV57DpoZaXGHhtju8Yztks8aXEaLVSXKDCR2i8/F89T9+D4dUl0U+C08wlcfG2tbe5aUY75X+J+cRpGKJJyh5u3xXfL1GprninlY9uyAc/Dk7Fl7ALAiovHN/4fhLv3jXFlNYfrlUdxffM+UDil5Z5nK7Wfi2VZnPn5HuanB8rc59pu8TxwQkrFDuzLJ+62y6KrHOXfOR2zc6+jqFRERA5nbWaQ6ctzeXN9PoHi56/8oaWb8b0SGZxWc0dsqEls7bBsT4CnV+by3u8FBA/6Ouvd0Mm13RM4p50Xt71mfp3J0VFgIrWasXMr3kf+im3HZqDwav3lNxM66cwYV1b9bGt+wfv4XRi5kd4tZmIKvpvuq9PL0tYm9l+X4HliStEKR43SKLjlQawWbWNbWE0TDOCdOhH7+lUAhI49Ad9N94Pt6K+y5QRNrpuzj483+w65X8t4G2+PaES3BuXvYO+a+SKuD14BINRvKL7x9x5VrSIiUjrLsvh+V4DHfs3l8y3Ff547bXBB+zjG90yo0M/wWFOT2JonaFp8vKmAZ1bmsWhX8YssdgPOauPl2u7xDGhScwM5qRwKTKTWMNK34pzzKcaedKxGTTGbt8H9vycw8nIAsBKSKBh/L2bX3rEtNIYiAdId2HZsASLTGvzjJhMaqD4KseSY9THulx/GMCOXJcLtuuK76T6slIYxrqxmMvbuIm7KVRg5WQD4zx1L8I+XH9Ux12UFueybDNZmFS3xZxCZgrP/l+CBTWATnQbPDWvA6a28h683Yzdxt4/GCPix7A7yp75UN/sm1SJVtWS0xJY+r/Vb2LT4ZLOP6ctzWLK7eKiQ5DT4S5d4rumeQPP42j0d4sfdAZ5bncfMcjaJ1ffFkSvtY9fAbfDS2nxeWJXHtvzin4AGboM/d45nXNd4WiboY1xfKDCRWsEx5zPcL04rWvsTCyyL/XluuHlbfDffj9WkeeyKrCnycvA8MQXHyh+jmwIjzgWXJxo2BYeOxGqqE7oqZ5q43nke1yf/i24K9T0R37V3gdsTw8JqPvuKpXim3YZhmViGge///kW4V/8jOtbHmwq4bu4+coKRX3dJLoPnh6bSMdnBq2uL/lA6taWbyYuyoysqGMCU45KY2CvhkFeP3M89gHPe50DhdMDLbjyiOmuTmvwHelUuGS2xo89r/VUQsnj9t3yeWJ7DhpziJ7DN42xc1yOBKzrHk+SqW/0+ytMktlWCnUd+zdX3xREo7WdK2AKHAaGDzo67pTi4tnsCF3RQX5n6SIGJ1HhG+lbiJl+OYZml3h/q2hvfTfeBV78YokIh3C8/jHPOp9FNFkbkt4ERuaTuHzdJjUarUimrGAX+cEGkt46tdl/9qi7Oj2bgfud5IDKCLP+eZ7EaNS3348OmxdRlOfz7l5zotu4pDmac0rDMZYPzgiY3zMvk/Y0F0W0XtPfy+OAGeB0lQxPbpnV4p1yNYVlYcQnkTftfpfZcqYmq68Q1bFoUhC0KQpG3vpBFfsjCF478yw8Vbiu8ryBssS03zH9W5ZXa0NcA3vtDQ45v4tIfvLVMVS4FLjVDWVf6n1+dx7Or8tjjK/43YPcGDsb3TOS8dl5cdbxvxOGaxJZG3xeHdqifKfsZwOmtPFzbPYGhzTTtpj5TYCI1nuutZ3F++kapgYllGATPuJjARdfEoLIazrJwvfWfyMfu4LsADBv5D76iqQNVIScT72N3YV+3HADLsBEYPZ7gqefEuLBaxjTxPHYXjp8WAJGpTAV3Pg7Owy/Tl+k3uXJ2Bl9v80e3ndvOy+ODU0hwHvpk2bIspv2cw/3LioKWPo2cvDa8YfGh3paF51//Fx3N5b/keoKnX1iRV1jrHOqPTAOYdkIyyS5bNOjw7Q88QlbJbQeFIQffPriBY2VKdhk0i7PTNM5OU6+N5vF2mnojtyPbbTSNs+NUr4CYsyyLiQsymbE2n7K+JI5NdXJZpzi6pDjonOKkqdemk5tapLQQ1rQi/UgO/jkwtJmbCT0TOKWFu15+jg/VJPZAdgMm9ExgSr/kaqyu9rh7SRZPLM8t82dKn4ZOXjwplXYKnATQV4HUTKaJbf1KHEvn4Pz2gzJHl4CBsXdntZZWaxgGGLbCESXFf6kagGWAc/anBC68Ojb11VFG+la8D9+Obec2ACyXB98NdxPuPSjGldVCNhu+q+8gbso12HZvx/77atyvPYH/z7cc8mHLM4KM/nYvGwuHbtsMuKdfEjf2OPTUmv0Mw+C23kl0a+Dk2jn7yAtZLNsTZPhHu5hxSkP6NY4ENvZfFkXDErNxc4Kn/OnoXm8tMGNdXokAdj8LuPX7rOos54hlBSyyAiFWZ4YOuV9jj60wRLFFw5RooOK10zzeTiOPDVsFTtxq8nSmmmKPL8ys7X6+3eZn1nYf2/MPnZ79nBHk50VFX3tJToPOKQ46JTvpkuygc4qDLslO2iTacSgEq1HWZ4WYMD+z1BB2f1iyf7n38T0T6N3o8IF5XdY20cG9/ZO5o08SZ362mx9KaRALkcDpl4zS76vPcoImL6/J47lVeWWGJTYD2ic5FJZIlL4SpOYww9jX/IJ96RwcS+diy9xz+McYVGiIfn1j7EnngMYvxVn775fKYlv7C95H78LIK1ypKKUhvpunYrbtHOPKarH4RHzj78H7jxswggGc331IuGMPQif+odTd39mQz/h5mRSEI1/zDd02XjwplWHN3RV+6rPaeGk3ysEl3+xlS26Y9AKTUZ/t5rFBDbi4nQv3G09H9/VfeHW5Rr7UdmuyQoQreVyqAcQ5DDx2A68j8s9jN/AW3vY4it4/8K3HccD7dohz2HhvYz4fbfKVOQKmc7KDVI+N9PwwO/LD+MIl9zvQbp/Jbp/Jrxll7+MwIM1bNCql+f6RK3G26CiW5nF2kl0Gr/2WX+JK+mPLc+t9vwFfyGLRLj/fFYYkR3uilx20WLo7yNKDGoO6bNAhKRKgdE520iXFQafkyD9N06o++SGTX/cG+XFPkJfX5h1yWkTvhk5eOjmVtok6ZTmQ12EwtJmbn/YGS/2ZbAHfbPNz9ud7uLFHAqe2dFco2K1rdhWE+c/KXJ5fnUdW4NC/xAygdYKmTksR/fSR2AqFsK9ahmPpHOw/zsOWva/ELpbNBoWrixz4o35/79fgsJHVUmptZDVqWmZeorCpcjkWfoP7+QcwQpE/0MMt2+O75QGshk1iXFntZ7bphP+Km/E8/yAA7pcfxmzdEbN1h+g+IdPi7qVZPLUiL7qtTyMnr5ycSquj6GTfM9XJd2c15vJvM1iwM4A/DNfO3Ufy3HlcsH0TAOGOPQj3H3bEz1EbhE2L/67J4+utZS/JbACDmro4q423KOw4IASJBBtGJBw5IPxw2ai0ofW9Up18tKn0Gg0DXj+1qH+NZVlkBSx25IejAUp6gcmO/DA78sKkF4RJzzdJzw+XaAB4oJAF2/LDhasplH2i7yplesF+4+dlMjDNXW/6DViWxarMEN9u8zFru5/56YFoyHkwr92gTyMHC3cGS/1VZgOeG9aA3KDFmqwg6zJDrMkKsTm3ZBoWMGFVZohVmSGg6OvEAFol2OmS7KBT4WiUyKgUB6me8p84afRQSYGwxYp9QZbtCfLjngDL9gRYnVm+4NVmRAIuhSWlG90pnseW5x5ynzk7/MzZ4adLsoMbeiZwYfs4PKX046qr1meFeGJFDv/7reSKQ2WxgDGd62+ALSWph4lUv2AA+4qlOJbMwbFsfnRZ4ANZDifhnv0I9R9GqPcgHMvm435hWtH0ksIQQI1LD21/w1wss2TYZBjkP/iqepgcLcvC+fFr0eakAKGe/fHd+Hc1Iq5k7hf/jXP2xwCYTZqT//f/QHwiuwvC/GVWBvPSA9F9L+sUx0MnpFTaH4aBsMXtizL575p8EkP5rF70f6QFIyOJ8v/2JGbHHpXyPDXRL3sD3Lwgs8yh3/vVlCaDr63LY3wlNqU1LYs9PrMwWDGj4UpR0GKSXhBmd4FZrmaMZWngNhjWzEOPBg56pDrp0cBJ6wR7nenTsKtg/zSbSEiSXlD2NJtjUp0Mb+Hm5OYeBjRx4XEYFf685odM1mWFWJsZYm1WiLVZQdZmhvgtO0SwAv1xGnlsdEp2FE7tKRqV0jLeXuyKvVbxiQTXqzNDLNsTYNmeIMv2BliRETzifkTqw3F4pX5fWJFm5Ut2B0qsKtTYY+PKbpFlcRtVIAysbZbtCfDorzl8uNFX7Oey0wYXdohjQs8Elu4OVOrvCqm7FJhI9fD7sP+6GMfSOTh+WohRkFdiF8vlJnzMAEL9hhHqfUKJk01j51acsz8tWhp32Eid7JeDY+5nRWGTaWIU/uoIHjcE/4R/xLi6Wq6U1YiCw87Ef/lN4NAVsUoX8OO9bzz2jWsBCPUdzPyL/8blszILr+5H/hh6YEAyY7vEV/qJpmVZvLA6j4LXn+OOTR8A8FnzgTSf/I86eRU5N2jywLIcnl6ZW+xq8AlpLhbvDOxfcKtG/pG5ITtUbMnoMZ3jqzzICZoWOw8YpVIUrERClh92B8gOVuxPriSnQfcGzmiA0qOBg+6pThIP07i4JigIWXy/s3CazXY/yw8xzaZ5nI2TmnsY3sLNsGZuGntLP5GrjM9ryLTYmFMYohSORlmbGWRtVii69Hh5xDmMaJDSyGPj6ZWlr85UU4LEymZaFr9lhVi2NxgNSH7ZGyxzpNB+dgO6pjjo08hF30YuGnsMrvhuX6n9JOrqx66ylfV9ETYtPtvi48kVuSzcGSj2GI8dLukYx/U9EuiU7IxR5ZXLsiy+3e7n0V9ymJte/PUmOg3+3CWea7sn0OKA5u2x+F0htY8CE6k6Bfk4fvkex5LZ2H9ehBEoOUza8ngJHTuQUP+hhI8ZAG5vDAqt+/aHTbZtv2P/aWGkA31KI/IfekMn9kcqPxfPE1NwrPghusl/4dUER14SCaekShi7dxA35Zpon5i7OlzMA63OAqCp18bLJ6cyIK3i/UrK/fx7d+G5bTT2UAC/4aDn8dPISE7jpZNSObmFp8qet7p9trmASd9nsTWv6Opk52QHjwxKYXBTt/7IPAL3LM3i8eW5ZU5FsBuUuz9M20R7JEApDFJ6NnDSNtGOPYYNTS3LYsW+EN9t8/Htdj8Ld/rL7BET5zA4samLk5t7OLmFmy7JjpiPpLEsi/QCk7WZQdYUjkpZkxlkXVbokKNhDscgEjJe2D6OJl4bTbz26NvSliqviSzLYlNuOBqM/LgnwM97g4cNmAygU7KD3o2c9G3kok9DJ70aOkv0i6nsUWFS0g+7AzyxPJcPNhWU6BlzeisPN/ZMYHBa7Vw6N2RavPd7AY8tzy0RzDbx2riuewJ/6RJPirvmB81SMykwkcqVl4Pjp4WRkGT5YoxgyStKVlwCoT6DCfUbSrhnP3BV3cmNlOR57E4cP84HwHf9FEIDTo5xRbWPsScdz8OTsW/bCIDldOK/6g5CA4bHtrB6wvxpEQmPTMaGRRiD04+9g4IufXjp5FSaxlXtEGP3f+7HueBLAF7qeCZXtrwEiFwJva9/Mtd2r/yRLdVpW16Y27/P5OPNRQG32w63HpPIhF6JuO2197XF2qGWZLYZ8P2fmmABKzKCrNgXZPm+ECsygsVCq0OJcxh0SymazrP/bYMqPElIzy+cZrM9Ms1mVxnBggH0buTk5OaRaTbHN3HVqq+lTH9kes+awmk9awtHpWzMDR+yYenhJDmNYgHKgW/TDrjd2GPDVQkfr/L0WLEsix35Jj/uCfBT4bSaZXuCZPgPHxq1TbTTp6GLvo2c9G7k4tiGTpJc5fv6UwhbPTblhPjPqlxeWZNP7kHNmXo3dHJjzwT+2NZbK5ZUzwuazFiXzxMrctlyUM+iDkl2xvdM5OIO9atni1QNBSZy9HIycfw4PxKSrPwRI1xymUYrIYnQcUMiIUn3vuCoG8P/aiP7ih/w/uv/AAh37kXBndNjXFHNZ6RvxTknMh0Mpwv7Twux5UaWsLQSkiiYeB9m514xrrJ+2J4X5orv9nLG4je4e9NMAHI8yYT++RzOxlXbYNe2cS1xUyLLcFvxSaTf9ypX/RDi8y1F4cKYTnH8e2BKrToZhMgVumdX5XH/j9nF/og+ubmbhwam6MSlkhzJlfRMv8nKfZEQZX+YsnJfiLxDdaI9QIs4Oz1SHcVClI7JjjJPiA51Up0fMlm4M8C32/x8t93Hyn1lL8vcMt5eGJC4GdbcTcM62C/BF7JYnx3i/mXZfLrZd1Q9bA6ngdsgrTA8SYsrfFtKyNLQYyt16eSyeqxMPT6J9knOwoaskek1O8sxoqZFnL1o5EgjJ70bOivUIFdiKytg8sqaPJ5ZmRed0rpfy3g713SL5/Iu8SSXM/CqTnt9YZ5dFVka+OAgr28jJxN7JXJma09MR9xJ3aLARI6IkbkX+w/zcCydjX31TxhmyV+uZnJqJCDpN5Rwl2PArj+4awTLwvvXv2DfvhGA/Hufw2zTKbY11WCOOZ/hfnFaUSc1y4o20DXTWlLwfw+ol041WZDu58+zMthVYGJYJp8sn8Zpe38BIivVFNzxaNWFsZaF54Gbcaz+CQD/ZTcSPO18TMvivh+zeeiXopUKTmji4pXhqTQpow9DTbNsT4CbFmTy896iEYGNPTamDkjmvHbeWj1ipiaqjCvppmWxKSfM8n1BlmcUBSm/55RvNIrLBl1SnNEGsz0bOOmZ6uSLrb6SJ9UWnN3Wwz6/xfe7/GWuNJHgMBjczM3w5m6Gt3DTMSn202yqyyFHDwEPD0rGZhjsKjDZWRBmV0GYXQUmuwqbBVe0t82hGEQa1TY+YJSKy2YwY13+EQc6jTy26KiRvo2c9GnoIq2KR/NJ9QiaFu//XsATK3KL/Q6ASN+PMZ3juLZ7Aq2PYrW5yrIpJ8QTK3KZsTa/RK+cU1u4mdgrkROb1s5pRVKzKTCRYg68km41akpw6EisppGTQWPvLhw/zMGxZDa2dcsxrJJfOmZqE0L9hhLqPxSzY0+w1bxkWsDxzQd4XnkEgODQkfjH3Rbjimqm/asMGVbJQNAC8qc8g9W+a/UXVs9YlsV/VuVx1+Ks6PKurRLsvH68g+MfvwHbnp0ABEacS2D0hCqpwb5sAd5H/wqAmdaC/PtfKhbOvLMhnxvn7Yv2bGgZb+e1U1I5tqGrSuqpDNkBk3/+mM3zq/OKnej9pUscU45L1nzvWig3aLJqXyg6GmV54ciU7EDl/6lnELmau78PSf/GrkqZNlJbHU0fjvyQya4Ck92FgcruaLBiRsOV/bcP11T1aCW7DPoU9hvpUzh6pGV83VmtSUpnWRbz0gM8sSKXL7YU7zloN+CPbb3c2COBvo2r/3farxlBHv81h5m/FxTr82Q34Lx2Xib0SqRnqkauS9VRYCJRxa+kE720FOo/DNuendg3rCr1cWbj5pHlf/sNxWzfVQ0vawNfPvE3XYBRkIfldJH36NuQoGX7DuZ661mcn71R6ggqy2YjeMbFBC68OgaV1R/5IZObFmTy1vqC6LaTmrt5YVgDGnrs2H5fjfef4zFCkStjvmv/RmjgKZVbRChE3F1/wbZjCwAF4+8l3G9oid2W7Qlw2Td72Z4f+XqJcxg8dWID/tSuZjWztiyLDzf5mLwokx35RV/b3VMiTV2rsmmuVD/LstiaFy4MUYrClHXZoQr332iVYGd4YR+SYc3dVdofpTaq6j4clmWRG7JKBCo7C0x2F749cPRKeZfz7ZTkYHKfRPo0ctEuUeFIfbc2M8hTK3J5Y31+icbNA9Nc3NgjgTNae4otq13ZLMtizo4Ajy/P4Ztt/mL3xTkMxnSK44aeNWPki9R9CkwEOPSV9NKYzVoXhSStOyokqYVcr03H9eW7APgvvIbgqEtiXFHN437qXhyLvyt1NJVl2AgdfxL+6++OQWX1w8acEGO+zeDXA7re39Qrgbv6JhWbo+/47iM8Lz0EgOXyUDDlacyW7SqtDsc37+N55VGgsO/PXx8v82deen6Y0d/uZenuoppv653I5N6JVfrHZXltyglx2/eZfLG16A9Qr91gcp9Eru+RUCsa/UnlKAhZrMks7I2yL8jb6wvY7Su7aetpLd28cWpDnUzXEpZlkRWw2FUQ5l8/Ra7Ol/bZtRswoWcCU/rpookUt8cX5vlVeTy/Oo89B/1s6JBk5/oeCVzSMa7EqkdHI2xafLTJx2PLc1i2p/gUoYZuG1d3j+eqrvHqlyPVSoGJAIe+kr5fuFWHaEhitWhbfcVJlTDStxJ/+2gAzEZp5E/7H9j0C+hArreexfnp66UHJhphUqW+3eZj3OwM9vkjH/t4h8FTQxrwx7aljNawLNzPP4hz3ucAmE1bkf/3Z8BbCctR5ucSf9tlGDmRJr/5dz+F2aH7IR/iC1nctGAfbxwwKubM1h6eGdqABGdsrsgHTYunV+TywE855B/QLPS0lm6mnZBCm0RdpavvDrXssU6qa7fDrdC09Nw0NXaWMhWELN5an8+TK3JZm1W80XOq28bYrpEQ42j62vhCFq//ls/05TlsOKgXU+sEO+N7JnBZp8oNZ0TKS191grF3J47F30EZYYllGISOPYGCf75A8I+XKyypI6ymLQkdMwAgMuVq2YIYV1TzBI8fFul4eBCr8L/gsJHVXlNdZ1kWD/+Sw3lf7o2GJR2S7Hx9ZuPSwxIAw8B/xc2EW3cAwJa+Bc/zD5b6uaso18f/i4YlwQHDDxuWAHgcBk8PacA/+yexf8DGx5t9nPbJbjbllL2qSFVZvMvPsA93cffS7GhY0izOxssnp/LmqQ0VlggAozvFl9kU1ALGdK6EAFJiokOyg+mDU7AZkfDrwLfTB2sVLDk0r8Pgii7xfH9OE948tSFDmhb1Mcnwm/z75xx6vZ3OjfP2sWpf8BBHKinTb/JQ4eNvXphZLCzplerk+WEN+PG8NK7qlqCwRGJGI0zqMSN7H86PZuD89sPo/P/S6Ep63WX/eRHeh28HINS9L77bH45xRTWL8+PXcL/9HAAWRmQaRmGPH/+4SYSGnBHbAuuYnKDJ9XP38dGmooZzp7fy8MyQBuVqQGrs3Ebc36/GyM8DwH/J9QRPv/CI6zH2pBM3eQxGMIjlcJL/wCtYjZtV6Bhfb/UxdnZGtOlmQ7eNV4anMrhp1fcJyfSb3PNDFi+tKVodwwCu6hbPXX2TSKqBy0VKbB1N41Kp+aq6x4rUHz/vDfDkilxmbijg4BXOT23h5oYeCZzU3I1hGKUuVe5xGDy1IpeX1+QVW8oeYGgzNzf1SuDkwseLxJoCk/ooLwfX52/h/OJtDH/RicmBf1AX22bYyH/wFS2dWheZJnGTL8e2cysA+ff9t1J7P9Rqfh9x/3cxtpxMLAyCJ50ZaZLbqCnBYSP1/VDJ1mUFGf1NBmsKh/sawOQ+iUw6tmK9Pw5czcay2Si4/RHMrsceUU3uZ/6Jc+HXAARGXkzgomuP6DjrsoJc/PVe1mdHrpw5DPj3wBT+3KVqTkAty+Ld3wv46+IsdhUUjRw8JtXJo4NSYrLKgdQeOqkWkfLalhfm2ZW5/HdtXonVuHo0cNC3kZPX1hVgGEUhbNiKTHE4cFy7zYCz23iZ2CuBPo30O0pqFgUm9Ynfh/Prmbg+eR0jLye62XJ5CJ52HmaDRrhnTI9cRbcsXUmvJ5xfvoP7tScACJ58Nv4/3xLjimoG51czcc94HIhMxVBz16rzyaYCrp27j5xg5NdRksvguaGp/KGV54iO53rneVwfzQDATE6l4N7nsFIaVugYtg2ribsnEpBYCUnk/es1iE88onogMtpj7KwMvt1e1Gz1qq7x3D8guVIbrf6eHeKWhZl8d8DzxDsM7uybxNXd4os1yxUREakMOUGTGWvzeXplLptzw4d/QCG3HS7rGM+NPRMUzEqNpcCkPggFccz+BNcHr2DLyohutuwOgiefRfCs0dGTCWPnVpyzP8XYk64r6fVFfi7xN52P4fdhuTyRJYaP4sSwTggFiZt0KbaM3QDk/+MFzML+GFJ5wqbF1J9y+PfPRQFutxQHM4Y3pEPyUfzhZIbxTJuEY+WPkefpfAwFtz8MjnIe07LwTr0J+5qfAfCPnkBwxLlHXk+hkGkxZWk2T67IjW4b2szNSyc1OOqO/4GwxePLc/n3z9nFloEc1drDgwOSaamlF0VEpIqFTIuPN/l4YkVOsdXiSnN8YxczTkmliVcLDkjNpsCkLjPDOBZ+g+u9/2LbvSO62TJshAaPIPCnP1d4Pr7UTa5XHsX1zfsA+C+5geDpF8S2oBhzzPkUzwv/AiDUexC+m++PcUV1T6bf5KrZGXy1rWgkxDltvUw/MaVyVpLJziRuylXR0Ctw+oUELrm+XA+1/zAX7+N/A8BMa0n+/S+VP2wphxnr8rhlQSaBwvHIbRPtvH5KQ7o1cB7R8ean+7llQWZ0OhNAy3g7/zohmZGty2iUKyIiUoX+9PluZu0IlHqfzYj8zn/hpNRqrkqk4tTxrS6yrMgf/HeNw/Ps/cXCklC/oeTf9yL+q+5QWCJRwVPPib7v/Oa9MldMqhfMMK6P/xe9GTjrshgWUzctzwhy0ke7omGJzYB/9EvixZMqcdndpBR8N/wdyx4JOlyfv4V9yazDPy4Uwv3mf6I3/RddU6lhCURWI/no9EY08UZe68acMCM+3s1nmwsO88jiMnxhbpi3j1Gf7YmGJXYDbuiRwPfnNFFYIiIiMdOnkQt7GbNADSLLBYvUBhqjW8fYV/yA653nsW9YVWx7qGd/AuePw2zXNUaVSU1mNW9DqMdxOFb8gG3Xduy/LCLce2Csy4oJx5LZ0Sa4oW59MDv2iHFFh1Za9/mjms5Sxd7dkM/4+ZnR5W1T3Tb+e1IDhjU/sn4lh2J27EHg0htwv/oYAJ7nHyS/ZXusZq3LfIzzuw+jn/9wl2MJ9z2x0usCGJDm5tszG3PZtxn8vDdIbsji0m8y+NtxSdzcK+GQKwNYlsXrv+XztyXZ7PUXhZvHNXLyyKAUjmmohnkiIhJbozvF89jy3FLv01LlUpvU3L+qpUJs61fieuf56Jz9/cIdexA4/0rC3frEqDKpLYKnnotjxQ8AOL9+r34GJpaF86PXojeDZ42OYTGHN2NdHhMOWgL0seW5NWoJ0P2BzsacMJtyQ/y4p2hOc++GTl4ZnkrrKuyvETzlT9h+W4Fz4dcYvgI8j99NwZSnwBNXcue8HFzvvxS96b/kukgT7CrSMsHBZyMbccPcTN7bWIAF3PtDNiv3BZk+uAFeR8nnXpsZ5JaFmcxLLxrmnOQ0uPu4JP7SJR67mrqKiEgN0CHZwfTBKWUuVa4mr1Jb6Cu1lrNt3YDr3Rdw/Di/2PZwqw4Ezh9H+NiBVfoHv9Qd4d4nYDZqim1POo5fF2Okb8Fq2irWZVUr+88LsW9ZD0C4fTfC3fvGuKKyrc8KMWF+JmYpXajGz89kYJo75n+MHBjohA+q89KOcTw0MKXUUKBSGQb+v/wfti3rsW/9Hfv2jbhf/Df+6/5W4mej6+PXMHKzAQgOPLVaRuTFOWy8eFIDevzi5J8/Rp77nQ0F/JYV4v7jk/lqq4/NuWGax9kpMC1eWZMX7X0CkTng9w9IplmchjaLiEjNclmneAamubVUudRqavpaSxk7t+F6/yUcC7/GsIo+hWaT5gTOHUdowMlgU4saqRjnp2/gfvMZAAIjziMwenyMK6pGloX3HzdgX78SgIKJ9xHuOzjGRZXtnqVZPL48t0QQsZ/LBh6HgUGkR4gNA8OIvL9/W+RtZHvxbYXbD9hmlLKttP33H8sftvhhT+kd8g1g6blN6JB8ZE1Oj4SRvoW4Kddg+PKBkivfGLt3EDf5coxQEMvpJP/BGVgN06qtPogsr3z1nH3khYo+qbbCVd4P/jS3TrDz0MAURrSs/KlMIiIiIhKheK+WMfbtwfXBKzjmfIIRLlo70mzQiMAfryA05IxKb1Ao9Udw6Ehc7/0XI+DHOfczAueNA28pUxfqIPvqn6JhSbhluxo/Jen3nHCpo0v2C5gQCBy4Q83Jxm0GzFiXz5R+ydX2nFbTVviumox3+t0AuF5/knDbzpidekZuv/0cRigS8ARPu6DawxKAUW28fDnKwflf7WFHfmQYSWmf4790juO+AcnEORSKi4iIiFQlnVnXFrlZuD55HedXMzGCRXPXrYQkAmdeRvCUP4HLHbv6pG5ISCI08FScsz/B8OXjnP9FsRV06jLnRzOi7wfPHF2jR2gtSPcza7vvkBFIqtugoceOZYFpWVhETr6jby2wsIptK3o/sr9V4r7S968oC9icGz7sfpUt3G8ogZEX4/r0DYxwGM9jfyN0wsnYtm/GsWIpAGZiSkxXRuqR6uSPbbw8syqv1PttBqS4bQpLRERERKqBApOariAf55fv4PrsTYyCoj+gLY+X4OkXEjj9QvDWjOaOUjcETz0X5+xPgEjz1+Apf6rzfXBs61dFG96aTZoTOn5YjCsqXU7Q5J6l2Ty/uvST6f1sBnx9ZpNqmyNsHRDI7A9R/vFDNk+vLH3KUCyXEwycfyX2Dauwr/4ZW84+nF/NLHZ/uFf/mP9M3e0zsRmljy6B2IRNIiIiIvWRLlHVVAE/zi/eJm7SpbhnvhgNSyynk8DpF5L379cJnPOXmP9hL3WP2boD4S7HAmDbsRl7YZBQl7k+LhpdEhh1KdhrXpb81VYfA9/bVSwsaZdoxwbYC3uT7H9b3d3nDcPAZhg4bAYuu4HbbvCXLvFljj6J6XKCdgf+866M1mYU/ttfl2PhNxiFywrHSusEO2VFlLEMm0RERETqm5p3VlDfhUM45n2B6/2XsWXsim62bDZCQ0cR+OMYrNQmMSxQ6oPAiHPwrvkZAOdXMwn37BfjiqqObcuG6CpTZoNGhAafFuOKisvwhbljcRZvri+IbotzGPytbxJXd4tnU264Rnafr8nLCTp+WgiGDSyz2HYDsAxwzv6UwIVXx6Y4YHSneB5bnlvqfTENm0RERETqmdj/VS0RpoljyWxcM1/Elr6l2F3BE04hcM5fsJq2jFFxUt+E+56ImdoYW8Zu7D8vxNi1HatJ81iXVSWcn/wv+n5w5MXgdMWwmiKWZfH+xgImfZ/FHl/Rif1Jzd08OiiFtomRH9/tkxzV2jy1ImrqcoLGnvSy77QOc381qMlhk4iIiEh9or+6qpmRvhXnnE8x9qRjNWpKcMgZ2HZtw/XO89g3/1Zs31DvgQTOG4fZumOMqpV6y+4gOPyPuN95HsOycH7zPoFLro91VZXO2LkVx/ffAmAlJhMcNirGFUXsyA9z68JMPtnsi25Lchnc1z+Z0Z3iMGpRT5maGOhYjZoWJRAHMwrvj7GaGjaJiIiI1CdGZmZmzVlrso5zzPkM94vTiv+hbpkl5qqHux6L//yrostdisREdibxt1yAEQxixSWQ9+jb4PbGuqpK5X7x3zhnfwyA//wrCZ41Oqb1WJbFq+vyuWtJFtkHLAk8qrWHfw9MoVmceldUBiN9K3GTLy/x89cCMGzkP/gKVppG9ImIiIjUd2r6Wk2M9K24X5yGYZkYphl5e9Af6+G2nSm4dRoFkx9VWCKxl5RCaMBwAIz8XBwLv4lxQZXLyNiFY97nAFjeeILD/xjTejbmhPjTF3uZMD8zGpY09th4+eRUZgxPVVhSiaymLfGPmwSGDctmxzJsWDYbGDb84yYpLBERERERQFNyqo1zzqdlDgG3gFDfE/FP+EedX75VapfgqefinPcFEGn+Gho2qs58jTo/ewsjHAIgeOo5EJ8YkzrCpsV/VuXxzx+zyQ8V/YC4uIOX+49PJtWjoKQqhIacQbhzL5yzD5giOWykwhIRERERiVJgUk2MPemlz5eHyGoNTledORGVusNs14Vwh+7Y16/EvnUDtjU/Y3btHeuyjl52Js5ZHwFgudwETjs/JmWs2hdk/Px9LN0djG5rGW/n0UEpnNrSE5Oa6hMrrWVMV8MRERERkZqtwlNy1q9fz2mnncZxxx3H8OHDWb16dYl9TNPkrrvuYuDAgfTv358bb7yRQCAAwIoVKzjjjDPo378/gwYNYuLEifj9/uhjU1JSGDRoECeeeCInnngiCxYsOIqXV3NEmwyWpoY0GZS6a31WiHuWZjFuVgb3LM1ifVao3I8Njjg3+r7rq5lVUV61c335DkYg8nMneNKZkJRSrc8fCFv866dshn64q1hYclW3eBae00RhiYiIiIhIDVDhwOSmm27iiiuu4IcffmDixImMHz++xD6vvvoqK1asYPbs2SxevBiAZ555BgCPx8O0adNYsmQJc+fOJTs7myeeeKLY47/88kvmzZvHvHnzGDRo0JG8rhonOHQkWCUHmViF/wWHjYxBVVIfzFiXR//3dvL48lze21jA48tz6f/eTl5bl1eux4f6D8NMTgXA/uM8jL27qrLcqpefi/Ob9wCw7A6CZ1xUrU//4+4AJ320i/uX5RAsXC24U7KDz0Y2YtoJKSQ61VpKRERERKQmqNBf5rt37+bnn3/moosiJxhnn302mzZtYtOmTcX2W758OSeddBIulwvDMBgxYgRvvPEGAB06dKBnz0hDU7vdTt++fdm4cWMlvJSaTU0GJRbWZ4WYMD8T04KwRbG34+dnsiG7HCNNHE5CJ58FgGGaOL/9oIqrrlrObz7AyI+ERaET/4CV2qRanjc/ZHL3kixO/WQ3K/dFPu52A245JoG5ZzdhYJq7WuoQEREREZHyqVAPk23bttGsWTMcjsjDDMOgZcuWbN26lTZt2kT369OnDy+//DJjx47F7XYzc+ZMtmzZUuJ4eXl5vPLKK/z9738vtv3MM88kGAwybNgw7rzzTuLj4w9Zl8/nq8jLiJ3+J5Pfpgue+Z9j37uTcMM0fINPx2zSHGrLa5Ba5aVVeWXOBLMsmDhvL3/rHU+PFDt2W9k9dPyDTiP1wxkYZhjHrI/JPv2iSN+d2ibgI+6LtwCwDBs5p56HWQ3fe/N3Brl1cS6/55rRbT1T7DwyIIFeqQ4I+fGVf5aUiIiIiIgcIY+n/NPfK9z01TioMalllexkeskll7BlyxZGjhxJfHw8w4YNY86cOcX2CQaDjB07lpNPPplRo0ZFt//666+0atWKvLw8br75Zu6++24eeuihQ9a0fft2wuFwRV9K7PQbUfS+PwylhEkilWHNbheWZae0BjoWMHdniNO+yCLebnFskknvpDB9k026JZi4Dhp/ZnQ7jtQVi7HlZpH3xUwyjh1cLa+hMjVe/A22nCwA9nXvx6Yq/v7LDcH0jU5mpjuj21yGxZWtg4xpEcKRl8OW8s2MEhERERGRo2S322nfvn25969QYNKiRQu2b99OKBTC4XBgWRbbtm2jZcvi00kMw+D222/n9ttvB+Ddd9+la9eu0fuDwSB//vOfSUtL48EHHyz22FatWgEQHx/PlVdeyU033XTYupo3b16RlyFSb3TZm8fXew4/giIvbLBgn50F+yJL2Hrs0LehgwGNnQxs4uC4Rk5soy6GFZGeRC1+nkf8qItq18pOoSANlnwdvWmc9xdatWhVZU/31bYAt/2cx46ColElxzdy8NCABDolaalgEREREZGarkKBSePGjenVqxdvvvkml112GR9++CGtW7cuNh0HIlNkfD4fKSkp7N27l0ceeYQ777wTgFAoxNixY2nQoAGPPfZYsRErmZmZuFwu4uLiME2TmTNn0qtXr8PWVZEhNSL1yRltDaavKj0wsQG3HpvA2qwwC3b62XXAib0vDAt2hViwK8QjK8BhQO+GLXi7cQda7V6PY/NvxG1dj9mpZzW9kqPnmP0N9n17AAj1GYyzQzech3nMkdjjC3PHoize3lAQ3RbvMJhyXBJXdovHVptCJhERERGReszIzMwsOafmENatW8f1119PRkYGiYmJPP3003Tr1o3x48dzxhlnMHLkSHbt2sWoUaOw2+2Ew2Guu+46xo4dC8Bbb73F1VdfTY8ePaJhyQknnMC///1vFi9ezE033YRhGITDYY455hgefPBBGjRoUPmvXKSOC5sWZ3+xh/npkSW9DSIDQgwi03GmD07hsk6R/kCWZbEhO8z8nX4WpPtZuDPAptyS09zGpM/hv6v/A8BnLQfx4Zm3Maipi4FpbprG1eBRE+EQcXdcgW3nNgDy734Ks0P3Sn0Ky7KY+XsBt32fxV5/Ufh0Sgs3jwxKoXVChWdAioiIiIhIDFU4MBGR2uHBn7KZuiwHgKZeG+e087KrwKR1gp0xneNpn3ToE/ituSEW7gywYGckQFmdGcIdDrDx+wk0DuYQNOy0P+ExdrgjgWb7RDuDmroZlOZiUFM3bRLsJXoexYrj+2/wPP0PAELd++K7/eFKPf62vDD/tzCTz7cUjeZJcRlMHZDCxR28NebjICIiIiIi5afARKQOWpDu58zP92BaYDPg49MbMajp0S1bu8cXZuHOAKnvv8AffnwHgHvbnMu97c4rdf/mcbbCAMXNoKYuuiQ7YhMcmCbev12JfesGAApuf5hw976Vc2jL4pW1+dy9JIvsYNGP0j+29TDthBSaeGvwqBsRERERETkkBSYidcw+v8mQD3axNS8ypeaOPonc3jup0o5v7N1F3K0XY5gmvoQG3HvFc8zdY/HD7gABs+zHpbptDCwcfTIozUWvVCeOwqWM12eFmLEuj825YVon2BndKZ4OyZUzhcW+bAHeR/8KQLhDdwr+9mSlNKvdkB1iwvx9zCuc8gSQ5rXx74EpnNXGe9THFxERERGR2NKkepE6xLIsbpy3LxqWDG7q4tZjEiv3ORo2Idz3RBxL5+DJ3cfd4Z8IjTyVgpDFD3sCLEz3s2BngMW7AuSFivLYDL/JJ5t9fLI5Mm0lwWEwIM1FnMPg400+bEakt4oBPLY8t1iPlSMv1sL10avRm4GzRh9RWHJgoNMy3o5hWPxnZT4F4aLXN7pTHP/sn0yK23aII4mIiIiISG2hwESkDnlhdV40kGjgNnh2aCp2W+VPgwmMOBfH0jkAOL+aSWjgqXgdBic2dXNi4dSfoGnx695gYSPZAAt3+skMFAUMuSGLb7b5o7fDB411u2FeJh9uLKCx147XYRBnN/A6Cv8Vvh/nOOBtsW024hwGSWuXYV+/KnL8Vh0I9x5Y4dc6Y10eE+ZnRprlWnDwIJrWCXYeH5zCSc21WpeIiIiISF2iKTkidcTyjCCnfLwLf+HiNv87JZWRratoaohl4b1rXLQvSP7fn8Fs1/WQDzEti9WZIRYUjkBZkO4nveAQc3gqwZc/3c/wzBUAXHPseL5oOfgQIUvkdvR9h0Fe0OTBn3Ip64fkpR29/OuEFBKcGlUiIiIiIlLXaISJSB2QFzQZNysjGpZc3S2+6sISAMMgOOJc7P/9NwDOr97Df/Udh3yIzTDo3sBJ9wZOruwWmT508dd7+XKrv8xA4mgMyFoXDUvWedP4b8rxmHkll0o+UjYD0rx2hSUiIiIiInWUAhOROuCOxVmsyQoB0DPVyb39kqv8OUMDT8V66z8YeTk4Fn1L4OJrsZIalPvxRmGA8vU2f4npOAB2A67oHMfV3RMoCFnkh6yit2Gr2LaCkEV+2Cy27bavP44ea0aXP9E6yRndNy9klfqcFbU5t/ICGBERERERqVkUmIjUcjM35PPK2nwA4hwG/z2pAR5HNSzf6/YQHDoS12dvYoSCOGZ9TPDsMRU6xOhO8Ty2PLfU+yzgxp6JtE+q+I8p2+b1xL28FAAztTGTxp/PJIez2D5B86DApTCI2b/NF7b437p8vtrqK9G3BCLNaVsnaNlgEREREZG6SoGJSC22MSfETQsyo7f/dUIynZKdZT+gkgVP+RPOz9/CsCyc335AcOQl4Cj/j5UOyQ6mD05h/P6mqhB9O31wyhGFJQDOj18rqvGMi8FR8mPitBkkuwySXWUfp2cDJ19t81HanCELGNP5KFfxERERERGRGkuT70VqqaBpceXsDLKDkbP589t7uaxjXLXWYDVuRrj3IABs+/Zg/3FehY9xWad4lp6bxoSeCZzT1suEngksPTftiJcUNtK34lg8CwAzMYXgsFFHdBwoCnRsRmSK0IFvjybQERERERGRmk9/7YvUUvf/mM3S3UEA2ibaeXhgCoZRDVNxDhIccQ6OZfMBcH01k4LjT6rwMdonOZhSSX1XXJ/8D8OKTKIJ/uECcB/dcr+XdYpnYJqbV9fmsTk3TOsEO2M6xyssERERERGp4/QXv0gtNGu7j0d/jfT+cBjwwrBUklyxGTAW7n4cZvM22LZvwr72F2ybf8Ns3TEmtRh7d+GY/yUAVlw8wVP+WCnHrcxAR0REREREagdNyRGpZXYXhLlmzr5oW427j0viuMaHaMRR1QyDwKnnRG86v34vZqU4P3sTIxxZLSh4yjkQlxCzWkREREREpHZTYCJSi5iWxXVz97GzIDLl5JQWbm7sGftQIDT4NCxvpOeIY+HXkJtd7TUY2ftwzo4sJWy5PAT+cH611yAiIiIiInWHAhORWuTJFbl8vc0PQBOvjaeHNMAWg74lJXjiCJ54OgBGwI9zzqfVXoLzi3cwApGPTfCkMyExpdprEBERERGRukOBiUgtsWxPgHt/KBq58cyQBjTx2mNYUXHBU/8Ufd/5zftghqvvyfNyIs8JWHYHwTMuqr7nFhERERGROkmBiUgtkB0wGTsrg2BkJg4TeyYwvMXRrf5S2aymrQj1Oh4A25507D8trLbndn7zPkZBHgChE0/HSm1cbc8tIiIiIiJ1kwITkRrOsixuXZjJ7zmRERvHNXJy13FJMa6qdMER50bfd341s3qe1F+A68t3ALAMG4FRl1TP84qIiIiISJ2mwESkhnv9t3ze2lAAQJLT4IWTUnHaakDfklKEex2PmdYCAMfKHzG2bazy53TO+hgjJwuA0AnDsQqfX0RERERE5GgoMBGpwX7LCjLp+6zo7UcGpdA20RHDig7DZiN4yp+iN11VvcRwMIDzszeLbp55adU+n4iIiIiI1BsKTERqKH/YYuysfeSFLABGd4rjvPZxMa7q8IJDzsByR/qrOOZ/Afm5VfZcjvlfYtu3B4BQ38GYLdtX2XOJiIiIiEj9osBEpIaasjSLXzKCAHROdvDggOQYV1ROcQmEBv8BAMPvwzn3s6p5nnAI1yf/i94MnDm6ap5HRERERETqJQUmIjXQ51sKeGZlZNUXtx1eOCmVeGft+XYNnHpO9P3IEsNmpT+HY/EsbLu2AxDqcRxmh26V/hwiIiIiIlJ/1Z4zMJF6YntemOvnZkZv/6NfMr1SnbEr6AhYLdoS6t4XANvObdh/XVK5T2CaOD9+LXozeJZGl4iIiIiISOVSYCJSg4RNi6vnZJDhj4zIGNnaw1Xd4mNc1ZEptsTw15W7xLD9pwXYt/4OQLhjD8Jde1fq8UVERERERBSYiNQgD/+Sw7z0AAAt4uw8MTgFw6iZSwgfTrj3QMxGaQA4flmEkb61cg5sWbg+KhpdEjjrMqilHyMREREREam5FJiI1BDf7/TzwE85ANgMeHZYA1I99hhXdRRsdoLD/xS96fymcpYYtq/8AfuGVQCEW3UgfOzASjmuiIiIiIjIgRSYiNQAmX6TK2fvIxxZQZhJxyYyuKk7tkVVguCwkVhOFwDOuZ+DL/+oj+n86MDeJRpdIiIiIiIiVUOBiUiMWZbF+Pn72JoXBmBgmotJxybGuKpKkpBMaOCpABgFeTjmf3lUh7P9tgLHqmUAmGktCfUfdtQlioiIiIiIlEaBiUiM/XdNPh9t8gGQ4jJ4bmgDHLa6M2riwOavrq/fA8s64mO5PpoRfT9w5qVgq8VTlkREREREpEZTYCISQyv3Bfnr4szo7SdPbEDLBEfsCqoCZuuOhDsfA4Bt+ybsK384ouPYNv+G46eFkWOmNiE0aESl1SgiIiIiInIwBSYiMZIfMhk7KwNfZCYOV3WNZ1Qbb2yLqiKBA5cY/urImr86Pz6gd8nIi8HhPOq6REREREREyqLARCRG/rooi9WZIQB6NHDwj/7JMa6o6oT7nojZoBEA9p8WYOzeUaHHG+lbcCyeBYCZ1IDgsFGVXaKIiIiIiEgxCkxEYuD93wt4aW1kxZg4h8GLJ6XicdSdviUlOBwEh/8RAMOycH7zfoUe7vrkdYzC3ifBP5wPrtq/gpCIiIiIiNRsCkxEqtmmnBATFuyL3n5gQDJdUur+9JLQSWdiFU6jcc75FPy+cj3O2LsTx/wvALDi4gme8qeqKlFERERERCRKgYlINQqaFlfN3kd2IDJa4tx2XsZ0iotxVdXDSmpAaMBwAIy8HBwLvy7X45yfvYkRjjR6CZ56Lnjjq6xGERERERGR/RSYSK2xPivEPUuzGDcrg3uWZrE+KxTrkirsgWXZLN4dAKBNgp1HBqVgGHV4Ks5BgiPOib7vLMcSw0ZWBs5ZHwNguTwETjuvSusTERERERHZr26tXyp11ox1eUyYn4kBWIABPLY8l+mDU7isU+0YcTB7u5+Hf8kFwGHACyelkuyqX5ml2a4r4Q7dsa9fiX3LemxrfsHsemyZ+zu/eAcjGAmYgiefBYkp1VSpiIiIiIjUd/XrbE1qpfVZISbMz8S0IGxR7O34+ZlsyK75I032+MJcMyeD/eMp7uqbRL/GrpjWFCvBU4tGmbi+nln2jnk50eawlsNJ8PQLq7gyERERERGRIgpMpMabsS6PsiatWBZMXZZNQejQUztiybQsrp+7j/QCE4CTm7uZ0CshxlXFTuj4kzCTGwBg/2Euxt5dpe7n/Po9DF9kJaHQiadjpTauthpFREREREQUmEiNtzk3TFlxiAW8vaGAVjO2M/yjXUxelMnMDflsza05o06eXpnHl1v9ADT22HhmSANs9ahvSQkOJ6GTzgbAME2c331Ych9fPq4v3wHAMmwERl1SnRWKiIiIiIioh4nUfK0S7IfrDUrIgh/3BPlxT5BnyAOgRZyd/k1cHN/ExYAmLnqlOnHZqzeo+GlPgL8vzYrefmZoA9Li7NVaQ00UPPksnB/PwAiHcc76iMDZY8Dljt7vnPUJRm42AKEThmM1aR6rUkVEREREpJ5SYCI1XiBslTnCxADObuNhTVaI1ZnFR5Vsyw+zbWMB728sAMBjhz6NXPRvHAlRjm/ioom36sKLnKDJ2FkZBCMzcRjfM4FTWniq7PlqE6tBI0L9huFc9C1GThaOxd8ROvH0yJ3BAM7P3ozuGzzzshhVKSIiIiIi9ZkCE6nRZqzL46mVedHbNgCD6Go5B66Sk+k3WbI7wOJdkX8/7A6Qe0BvE18YFu4MsHBnILqtbaI9OgKlf2MX3Rs4cdgqZxTKrQsz2ZATBqBvIyd/65tUKcetK4IjzsW56FsAnF/NJDT4D2AYOOZ9gS1zDwChviditmwXyzJFRERERKSeUmAiNdYXW3xMnJ8ZvT2hZwJ2I9LTpHWCnTGd42mfVPQlnOK2MaKlhxEtI6M4wqbFyswQi3f5oyHK74UBxn4bc8JszCngrfWRUSgJDoPjGrvof0CIkuKueKufN37L583CYyY6DV4Yllrt04FqOrNjD8JtOmPftBb7xrXY1q/EbNcF1yevR/cJnDU6hhWKiIiIiEh9ZmRmZtbc5UWk3lqyK8DZn++hIBz58ry2ezxTj0/GOMpmqbsKwiwpDE8W7w6wbE8AX/jQj+mS7OD4JkUhSqdkR6lNW9dnhZixLo+V+4J8u81PsPA767mhDbigQ9xR1V1XOeZ+huf5BwEInnAK4WNPwPOf+wAI9eiH77Z/x7I8ERERERGpxxSYSI2zNjPIHz7dzT5/5Evz3HZenh9WNSvLBMIWv2YEWbQrEA1StuUfOkFJcRkH9EFxc1xjJ+9vLGDC/EwMIHzAd9QJTVx8PkrL4ZYp4Cf+lgsxcrKwDANcHgx/ZGRO/h2PYnbtHdv6RERERESk3lJgIjXKjvwwp32ymy25kdBiaDM3b49oiLsap7NszQ2xZHcgGqL8vDdI6BDfJfv7qZTGBiw9L63Y1CEpzvPwZBw/f19smwX4x91OaOgZsSlKRERERETqPQUmUmNk+k1Gfrablfsiq930SnXyyRmNSHJVvIdIZSoIWSzbE4iGKIt3BdjjM8v1WLsR6b0ypV9yFVdZOxnpW4mbPAbjoHWjLQDDRv6Dr2CltYxJbSIiIiIiUr/psrfUCL6QxWXf7o2GJW0S7LwzomHMwxIAr8NgUFM3g5q6AbAsi4054Wh48u7v+WQFSs8dLSJNaqV0zjmfgmHAQYGJAVgGOGd/SuDCq2NTnIiIiIiI1GsKTCTmwqbFNXMzmJ8eWe63odvGu6c1JC3OHuPKSmcYBu2SHLRLcnBxxziSXQaPL88t1rskui/QOqFmvo6awNiTXvZ8JqvwfhERERERkRiI/eV7qdcsy2Lyoiw+2OgDIM5h8NaIhnRMdsa4svIb3Sn+UOf8jOkcX53l1CpWo6aRVKk0RuH9IiIiIiIiMaDARGLqkV9zeW51HgAOA145OZXjGrtiXFXFdEh2MH1wCjYj0rPkwLfTB6eo4eshBIeOBKvkIBOr8L/gsJExqEpERERERERNXyWGZqzL48Z5mdHbTw9pwCUd42JX0FHakB3i1bV5bM4N0zrBzpjO8QpLysEx9zPcL0wr6mVSuOyQf9wkQkO0So6IiIiIiMSGAhOJiS+2+Lj0m73Rvh/39EtiYq/E2BYlMWPs3Ipz9qcYe9KxGjUlOGykVscREREREZGYUmAi1W7JrgBnf76HgsK05Nru8Uw9PhnDKKuZhYiIiIiIiEj1Ug8TqVZrM4Nc+HVRWHJuOy/3KywRERERERGRGkaBiVSbHflhzvtqL/v8kbBkaDM3Tw9pgE1hiYiIiIiIiNQwCkykWmQFTM7/cg9bcsMA9Ex1MmN4Km67whIRERERERGpeRSYSJXzhSwu/WYvK/aFAGidYOedEQ1JcunLT0RERERERGomnbFKlQqbFtfMzWB+egCAhm4bM09rSNM4e4wrExERERERESmbAhOpMpZlMXlxFh9s9AEQ5zB4a0RDOiY7Y1yZiIiIiIiIyKEpMJEq88ivuTy3Kg8AhwEvn5zKcY1dMa5KRERERERE5PAUmEiVmLEuj3t/yI7enn5iA0a09MSwIhEREREREZHyU2Aile6LLT4mzs+M3v77cUlc0jEudgWJiIiIiIiIVJACE6lUS3cH+PN3GYStyO1rusUzsVdCbIsSERERERERqSAFJlJp1mUFufCrvRQUpiXntPUydUAyhmHEuDIRERERERGRilFgIpViR36Yc7/cS4bfBGBIUxfPDG2ATWGJiIiIiIiI1EIKTOSoZQVMzv9yD1tywwD0THUy45SGuO0KS0RERERERKR2UmAiR8UXsrjsm72s2BcCoHWCnXdGNCTZpS8tERERERERqb10VitHLGxaXDM3g3npAQAaum3MPK0hTePsMa5MRERERERE5OgoMJEjYlkWkxdn8cFGHwBxDoM3RzSkY7IzxpWJiIiIiIiIHD0FJnJEHvk1l+dW5QFgN+Dlk1Pp19gV46pEREREREREKocCE6mw19blce8P2dHb0wenMKKlJ4YViYiIiIiIiFQuBSZSIV9s8TFhfmb09t+PS+LSTvGxK0hERERERESkClQ4MFm/fj2nnXYaxx13HMOHD2f16tUl9jFNk7vuuouBAwfSv39/brzxRgKBQPT+zz//nP79+9OnTx/GjBlDbm5u9L6lS5dy4oknctxxx3H22WeTnp5+hC9NKtvS3QH+/F0GYSty+5pu8UzslRDbokRERERERESqQIUDk5tuuokrrriCH374gYkTJzJ+/PgS+7z66qusWLGC2bNns3jxYgCeeeYZAHJzcxk/fjyvvfYay5Yto2nTpjz00ENApJHoVVddxdSpU/nhhx8YMWIEd95559G8Pqkk67KCXPjVXgoK05Jz2nqZOiAZwzBiXJmIiIiIiIhI5atQYLJ7925+/vlnLrroIgDOPvtsNm3axKZNm4rtt3z5ck466SRcLheGYTBixAjeeOMNAL7++mv69OlD586dARg3bhzvvPMOAMuWLcPtdjNkyBAA/vznP/PJJ58QDAaP7lXKUdmRH+bcL/eS4TcBGNLUxTNDG2BTWCIiIiIiIiJ1lKMiO2/bto1mzZrhcEQeZhgGLVu2ZOvWrbRp0ya6X58+fXj55ZcZO3YsbrebmTNnsmXLFgC2bNlCq1atovu2bt2aHTt2YJpmifsSExNJSEggPT292PaD+Xy+irwMqYDsgMl532SzJTcMQI8UOy8MjscK+vEpxxIREREREZFaxOMp/4IlFQpMgBJTMCzLKrHPJZdcwpYtWxg5ciTx8fEMGzaMOXPmlHmMih7/YNu3byccDh92P6kYvwkTV7hZmWUHoLnb5N+d88namUNWjGsTERERERERqQi73U779u3LvX+FApMWLVqwfft2QqEQDocDy7LYtm0bLVu2LLafYRjcfvvt3H777QC8++67dO3aFYBWrVoxd+7c6L6bN2+mWbNm2Gw2WrVqxebNm6P35eTkkJubS9OmTQ9ZV/PmzSvyMuQQNuSEeX29j815Jsv3hVifE5mGk+o2ePvUVDok2WNcoYiIiIiIiEjVq1Bg0rhxY3r16sWbb77JZZddxocffkjr1q2LTceByBQZn89HSkoKe/fu5ZFHHok2bz3llFO49dZbWbt2LZ07d+aFF17gvPPOA6B37974fD7mzp3LkCFDeOmllxg1ahROp/OQdVVkSI2Ubca6PCbMz8QATAv2j+1x2eCtEY3o0dgVy/JEREREREREqo2RmZl5+DkvB1i3bh3XX389GRkZJCYm8vTTT9OtWzfGjx/PGWecwciRI9m1axejRo3CbrcTDoe57rrrGDt2bPQYn376KVOmTCEUCtG9e3eefvppkpKSAFi8eDE33/z/7d1rjN31fT/495mL7RnbYEjMzIAvraM/JG6NkrKR0NReGnKRDNSVtosc4iLSFB7EkhcjBfUBCcsirVCScmmaBgltpagBItSKCiIQSioUYceV2lisN1vKGkFqwJ5zMASD7Zkxczn7YDzjc8b4cuzxnNvrJSEO5/zO6Pvjd5583/p+Pp+7Mjo6moGBgTz22GNOkMyD1z8Yz+f/pZTJj/k1FJLs/vO+rLmo5gouAAAAaEo1Bya0pv/j1x/kB//vkUx8zK+hs5D8b3+4JP/7/3Tx/C8MAAAA6qCmscK0rjePTORUyVn5+OcAAADQLgQmJElWLenMqWYXFY5/DgAAAO1CYEKS5C/+x+LTnjC59crF87kcAAAAqCuBCUmST13clb/742VV73UWko5C8nd/vEzDVwAAANqKpq/MKJfL6f/HAzk2mSztLuT2Ty/OrVcuFpYAAADQduyEmfHBR+Ucm5x6fc3yBabiAAAA0LaU5DCjNHJiEk5fj58GAAAA7cuumBnF4ROBSX+PqTgAAAC0L4EJM4ojkzOv+3sFJgAAALQvgQkzSpUnTHr9NAAAAGhfdsXMGKoKTJwwAQAAoH0JTJhRqizJ0cMEAACANiYwYUblCZM+JTkAAAC0MbtiZkz3MLloQSG9XX4aAAAAtC+7YpIk5XJ5piRHOQ4AAADtTmBCkuTwWDlHx8tJNHwFAAAAgQlJktJIxYScHj8LAAAA2pudMUmSoeGKCTlOmAAAANDmBCYkOdHwNUn6BCYAAAC0OYEJSZLisJIcAAAAmGZnTJKkOKIkBwAAAKYJTEgyu+mrwAQAAID2JjAhSTJU1cPEzwIAAID2ZmdMkqR0fErO0u5ClnT7WQAAANDe7IxJcqLpq/4lAAAAIDAhyZGxyRwZLydJ+kzIAQAAAIEJJ8pxEidMAAAAIBGYkGTIhBwAAACoIjAhJRNyAAAAoIrdMVUjhQeU5AAAAIDAhKQ0cqKHSZ+SHAAAABCYcGKkcJL0K8kBAAAAgQmzAxMnTAAAAEBgwkxJzuKuQpZ2+0kAAACA3TEzJ0yU4wAAAMAUO+Q2d3RsMh+OlZNo+AoAAADTBCZtrnJCjv4lAAAAMEVg0uZMyAEAAICT2SG3udJIRWCiJAcAAACSCEza3tCwkhwAAACYTWDS5koVJTmavgIAAMAUgUmbG6ooyRnQwwQAAACSCEzaXqmiJKdPSQ4AAAAkEZi0vemmrz2dhVzUXajzagAAAKAxCEza3NDxHib9vR0pFAQmAAAAkAhM2trIeDkffFROYkIOAAAAVBKYtLFSRcPXfhNyAAAAYIbApI0VK0cKm5ADAAAAM+yS21ixYkKOEyYAAABwgsCkjRUrS3L0MAEAAIAZApM2VhquDEz8FAAAAGCaXXIbGxp2wgQAAAA+jsCkjZVG9DABAACAjyMwaWPTU3IWdSYXLyjUeTUAAADQOAQmbWy66WtfT2cKBYEJAAAATBOYtKljE+W8f6ycRP8SAAAAmE1g0qaKJuQAAADAKdkpt6nSyInApE/DVwAAAKgiMGlTQ8MnJuQMKMkBAACAKgKTNlUarjxh4mcAAAAAleyU21RxpLKHiRMmAAAAUElg0qaKFSU5AhMAAACoJjBpU5VNX/uV5AAAAEAVO+U2NXS8h8mCjuSShX4GAAAAUMlOuU2Vjpfk9PV2plAo1Hk1AAAA0FgEJm3oo4ly3js2FZgM9OhfAgAAALMJTNpQZf+Svl4/AQAAAJjNbrkNlUYqJuQ4YQIAAAAnEZi0oemGr4mRwgAAAPBxBCZtqDSsJAcAAABOx265DRWHT5TkDDhhAgAAACcRmLShYmXTVz1MAAAA4CQCkzZUrOph4icAAAAAs9ktt6Hi8Sk53R3JpQv9BAAAAGA2u+U2NN30ta+nMx2FQp1XAwAAAI2nq9YvvP766/nmN7+Z9957LxdffHF+9KMf5dOf/nTVNeVyOffee29+8YtfpKOjI5deeml+8IMfZM2aNfnlL3+Zb3/72zPXvvvuu7nsssvy0ksvJUmWLVuWtWvXpqNjKsv53ve+l8HBwfO5RyqMTZZzcHTqhIlyHAAAAPh4NQcm27dvz2233ZYtW7bkmWeeybZt2/KLX/yi6prnn38+u3btyo4dO9Ld3Z3vf//7uf/++/PjH/84f/Inf5KdO3fOXLt58+asX7++6vs///nPs2TJknO8JU7nnZETE3I0fAUAAICPV9MRg4MHD2bPnj3ZvHlzkmTTpk3Zt29f9u3bd9K1x44dy+joaMrlcg4fPpzLL7/8pGuGhoby0ksv5atf/eo5Lp9aVTd8FZgAAADAx6nphMn+/fszMDCQrq6prxUKhaxYsSJvv/12Vq9ePXPdxo0bs3Pnzlx11VVZsmRJBgYG8txzz530937605/mS1/6UpYvX171/k033ZSxsbFcd911ueeee7J48eLTrmt0dLSW22hrb33w0czrT3ZP+n8HAABA21i0aNFZX1tzSU5hVpPQcrl80jV79uzJa6+9lldeeSUXXXRR7rvvvtx999159NFHq6574okn8sADD1S995vf/CYrV67M0aNHc9ddd+Xee+/Ngw8+eNo1HThwIBMTE6e9himvDnUlWZAk6R45lLfeeq++CwIAAIB50NnZmTVr1pz19TUFJldccUUOHDiQ8fHxdHV1pVwuZ//+/VmxYkXVdU8++WQ2bNiQZcuWJUluueWW3HzzzVXX/OpXv8rIyEi++MUvVr2/cuXKJMnixYtz++23Z/v27Wdc18eV+/Dxjr0/nGQkSfKZKz6ZlZcvqO+CAAAAoAHVFJgsX74869aty1NPPZUtW7bk2WefzapVq6rKcZJk9erVefHFF7N169Z0d3fnhRdeyNq1a6uuefzxx/O1r30tnZ0n+mgcOnQoCxYsSG9vbyYnJ/P0009n3bp1Z1xXLUdq2t17YyMzr1devCiLFglMAAAAYLaaS3IeeeSRbN26NQ899FCWLl06U2azbdu2bNy4MTfccEPuuOOO7N27N4ODg+nu7k5/f38efvjhmb9x+PDh/OxnP6ualpMke/fuzfbt21MoFDIxMZGrr7463/3ud8/zFqlU2fR1QNNXAAAA+FiFQ4cOndyEhJb1Pz/zTv6f342lq5C8c9vl6ZjVkwYAAACocawwza84MnXC5LKeDmEJAAAAnILApI2MT5ZzcGQySdKvHAcAAABOSWDSRg6OTma6/qqvR2ACAAAApyIwaSMavgIAAMDZEZi0kcrApK/XowcAAIBTsWtuI8XhyZnX/UpyAAAA4JQEJm1kekJOoukrAAAAnI7ApI2UKktyejx6AAAAOBW75jYyNHKiJEfTVwAAADg1gUkbmT5h0lFIPrnIowcAAIBTsWtuI9NTcvp6OtLZUajzagAAAKBxCUzaxMRkOe+MTpXk9JmQAwAAAKclMGkTB0cnM1meet2nfwkAAACclsCkTRQrJuQMmJADAAAAp2Xn3CZKFRNynDABAACA0xOYtImqEyYCEwAAADgtgUmbKI6cCEz6lOQAAADAadk5t4nKEyb9TpgAAADAaQlM2kRx+EQPE4EJAAAAnJ7ApE2UjpfkFJIsX+SxAwAAwOnYObeJ6ZKcy3o60tVRqPNqAAAAoLEJTNrAZLk8M1a4r0c5DgAAAJyJwKQNvDs6mYny1OuBXo8cAAAAzsTuuQ1UTsjp0/AVAAAAzkhg0gYqJ+QoyQEAAIAzE5i0geLIiRMmA06YAAAAwBkJTNpAqbIkp8cjBwAAgDOxe24DxZETJTlOmAAAAMCZCUzagKavAAAAUBuBSRuYDkwKSS5TkgMAAABnZPfcBkrHS3I+uagj3R2FOq8GAAAAGp/ApMVNlsspHZ+SoxwHAAAAzo7ApMX97thkxo73fB1QjgMAAABnxQ66xRWHT0zIccIEAAAAzo7ApMVVTsjp7xGYAAAAwNkQmLS44khFYNLrcQMAAMDZsINucUpyAAAAoHYCkxZXqijJGRCYAAAAwFkRmLS4ypKcPlNyAAAA4KzYQbe4yqavfZq+AgAAwFkRmLS44shUD5NPLOzIgs5CnVcDAAAAzUFg0sLK5fJMD5M+E3IAAADgrNlFt7D3j03mo+NDcjR8BQAAgLMnMGlh0+U4if4lAAAAUAuBSQsrVo0U9qgBAADgbNlFtzATcgAAAODcCExaWFVJjh4mAAAAcNYEJi1MSQ4AAACcG7voFlYaUZIDAAAA50Jg0sKKwydKcvqV5AAAAMBZE5i0sOmSnEsWFrKws1Dn1QAAAEDzEJi0qHK5nOLxkpwB5TgAAABQE4FJi/rgo3KOHW9hYkIOAAAA1EZg0qKGhisbvnrMAAAAUAs76RZVOSFnwAkTAAAAqInApEVVTshRkgMAAAC1EZi0qOKwEyYAAABwrgQmLao4oocJAAAAnCs76RZVWZLT74QJAAAA1ERg0qJKVSdMBCYAAABQC4FJi5ruYXLxgkJ6ugp1Xg0AAAA0F4FJCyqXyzMlORq+AgAAQO0EJi3ow7FyRibKSZTjAAAAwLkQmLSgypHCfb0eMQAAANTKbroFVU7IGXDCBAAAAGomMGlBxcoJOXqYAAAAQM0EJi2oVFGSM6AkBwAAAGpmN92Cqk6YKMkBAACAmglMWlBVDxMlOQAAAFAzgUkLMiUHAAAAzo/ddAuaDkwu6i6kt8sjBgAAgFrZTbeg0shUSU6/chwAAAA4JwKTFnN4bDJHx8tJkr4ejxcAAADOhR11iylWjRR2wgQAAADOhcCkxVROyOkTmAAAAMA5qTkwef311/OVr3wl11xzTa6//vq8+uqrJ11TLpfzne98J9dee20GBwdz00035Y033kiS7Nu3L5/4xCeyfv36mX9++9vfznz317/+ddavX59rrrkmmzZtSrFYPI/baz9VE3KU5AAAAMA5qXlHvX379tx2223ZvXt37rzzzmzbtu2ka55//vns2rUrO3bsyK5du3Ldddfl/vvvn/n84osvzs6dO2f++f3f//0kU0HLHXfckQceeCC7d+/Ol7/85dxzzz3ncXvtpziiJAcAAADOV02BycGDB7Nnz55s3rw5SbJp06bs27cv+/btO+naY8eOZXR0NOVyOYcPH87ll19+xr//8ssvZ+HChdmwYUOS5Otf/3qee+65jI2N1bLMtqYkBwAAAM5fVy0X79+/PwMDA+nqmvpaoVDIihUr8vbbb2f16tUz123cuDE7d+7MVVddlSVLlmRgYCDPPffczOeHDx/OF77whUxMTOTGG2/Mt771rXR2duatt97KypUrZ65bunRplixZkmKxWPX+bKOjo7XcRks7cOSjmdeXdo5ndLRcx9UAAABA41i0aNFZX1tTYJJMhSSVyuWTN+R79uzJa6+9lldeeSUXXXRR7rvvvtx999159NFH09/fn1deeSXLly/P+++/n7/8y7/MD3/4w9x5551n/fdnO3DgQCYmJs54XTt489DCJFMnS8beO5C3DtV1OQAAANAQOjs7s2bNmrO+vqbA5IorrsiBAwcyPj6erq6ulMvl7N+/PytWrKi67sknn8yGDRuybNmyJMktt9ySm2++OUmycOHCLF++PElyySWX5C/+4i/yT//0T7nzzjuzcuXKvPnmmzN/5/Dhwzly5Ej6+/tPu66zKfdpF4f2vJ9kMku6Crnq9059KgcAAAA4tZoCk+XLl2fdunV56qmnsmXLljz77LNZtWpVVTlOkqxevTovvvhitm7dmu7u7rzwwgtZu3Ztkqk+KMuWLUt3d3eOHTuWn/3sZ7n66quTJJ/97GczOjqaHTt2ZMOGDfnxj3+cG2+8Md3d3addVy1HalrdO8dLcPp7O/1/AQAAgHNUc0nOI488kq1bt+ahhx7K0qVL8+ijjyZJtm3blo0bN+aGG27IHXfckb1792ZwcDDd3d3p7+/Pww8/nCT5t3/7tzzwwAPp6OjIxMRENmzYkG9961tJko6Ojjz22GO56667Mjo6moGBgTz22GNzeLut7cjYZA6PTQUmfb1GCgMAAMC5Khw6dEhX0Bbx+gfjuebpUpLkf13Tk//rukvrvCIAAABoTo4htJDiyInGt309RgoDAADAuRKYtJDi8InApF9JDgAAAJwzu+oWUhyZnHnd74QJAAAAnDOBSQupPGHS1yswAQAAgHMlMGkhpYrAZEBJDgAAAJwzu+oWMjSs6SsAAADMBYFJCykd72GyuKuQpd2FOq8GAAAAmpfApIVMjxXu6+lIoSAwAQAAgHMlMGkRw+OT+fCjcpKkX8NXAAAAOC8CkxZRGq4YKSwwAQAAgPMiMGkR1Q1fPVYAAAA4H3bWLaI0UjlS2AkTAAAAOB8CkxZRrCjJ6ROYAAAAwHkRmLSIYkVJTn+PwAQAAADOh8CkRRQrSnL6ez1WAAAAOB921i2iqiTHCRMAAAA4LwKTFjHd9LWns5CLFxTqvBoAAABobgKTFjE9VrivtyOFgsAEAAAAzofApAWMjJfzwUflJEYKAwAAwFwQmLSAUkXDV/1LAAAA4PwJTFpA5Ujhvh6PFAAAAM6X3XULKI2cmJCjJAcAAADOn8CkBQxVnjARmAAAAMB5E5i0gFJFYDLQ65ECAADA+bK7bgHFipIcTV8BAADg/AlMWkCx6oSJwAQAAADOl8CkBUyX5CzsTC5eUKjzagAAAKD5CUxawNDIVGDS19OZQkFgAgAAAOdLYNLkjk2U8/6xchLlOAAAADBXBCZNrrJ/SV+PxwkAAABzwQ67yZVGTgQm/U6YAAAAwJwQmDS54vCJkcICEwAAAJgbApMmpyQHAAAA5p4ddpOrLMnR9BUAAADmhsCkyQ1VlOT09QhMAAAAYC4ITJpc9QkTjxMAAADmgh12k5vuYbKgI7lkoccJAAAAc8EOu8lNT8np6+1MoVCo82oAAACgNQhMmthHE+W8d2wqMOk3IQcAAADmjF12E6vsX6LhKwAAAMwdgUkTK42cmJBjpDAAAADMHYFJExsarjhhIjABAACAOSMwaWKlisCk30hhAAAAmDN22U2sWFGS06+HCQAAAMwZgUkTKyrJAQAAgAtCYNLEKktyBpTkAAAAwJyxy25iQ8dLcroKyaULPUoAAACYK3bZTWz6hEl/b2c6CoU6rwYAAABah8CkSY1NlvPu6NQJk74ejxEAAADmkp12k3pnZDLl46/7NXwFAACAOSUwaVKVDV8FJgAAADC3BCZNaqhypLCSHAAAAJhTdtpNqnR8Qk7ihAkAAADMNYFJkyqOVJTk9AhMAAAAYC4JTJpUsaqHiccIAAAAc8lOu0lp+goAAAAXjsCkSQ0NT/Uw6Swkn1zkMQIAAMBcstNuUqXjPUz6ejrSUSjUeTUAAADQWgQmTWh8spx3jk/J6VOOAwAAAHNOYNKEDo5Opnz8tQk5AAAAMPcEJk2oZEIOAAAAXFB2201oqCIw6XPCBAAAAOacwKQJlY73L0mSAT1MAAAAYM4JTJpQ1QkTJTkAAAAw5+y2m1BVDxMlOQAAADDnBCZNqFhRktOvJAcAAADmnMCkCRWPnzDpKCTLF3mEAAAAMNfstptQaWQqMLlsUUc6Owp1Xg0AAAC0HoFJk5mYLM9MyelTjgMAAAAXhMCkybw7OpnJ8tRr/UsAAADgwhCYNJmhqgk5Hh8AAABcCHbcTaZkQg4AAABccAKTJjPd8DVJ+nsEJgAAAHAhCEyaTGVJTl+vxwcAAAAXgh13kykNnyjJGVCSAwAAABdEzYHJ66+/nq985Su55pprcv311+fVV1896ZpyuZzvfOc7ufbaazM4OJibbropb7zxRpLkP//zP7Nx48Z8/vOfz+DgYO68884cO3Zs5rvLli3L4OBg1q9fn/Xr12fXrl3ncXutp+qEiZIcAAAAuCBqDky2b9+e2267Lbt3786dd96Zbdu2nXTN888/n127dmXHjh3ZtWtXrrvuutx///1JkkWLFuX73/9+/uM//iM7duzIhx9+mB/+8IdV3//5z3+enTt3ZufOnRkcHDzHW2tN0z1MCkkuMyUHAAAALoiadtwHDx7Mnj17snnz5iTJpk2bsm/fvuzbt++ka48dO5bR0dGUy+UcPnw4l19+eZLkU5/6VP7wD/8wSdLZ2Zk/+qM/yn//93+f5220j+mSnOU9HenqKNR5NQAAANCaumq5eP/+/RkYGEhX19TXCoVCVqxYkbfffjurV6+euW7jxo3ZuXNnrrrqqixZsiQDAwN57rnnTvp7R48ezT/+4z/mvvvuq3r/pptuytjYWK677rrcc889Wbx48WnXNTo6WsttNK3JcnnmhEnfokLb3DcAAADMhUWLFp31tTUFJslUSFKpXC6fdM2ePXvy2muv5ZVXXslFF12U++67L3fffXceffTRmWvGxsbyjW98I1/4whdy4403zrz/m9/8JitXrszRo0dz11135d57782DDz542jUdOHAgExMTp72mFfzuo2S83JskuSgf5a233qrzigAAAKA5dHZ2Zs2aNWd9fU2ByRVXXJEDBw5kfHw8XV1dKZfL2b9/f1asWFF13ZNPPpkNGzZk2bJlSZJbbrklN99888znY2Nj+frXv56+vr5897vfrfruypUrkySLFy/O7bffnu3bt59xXdPlPq3ug/fHk3yQJFl9SW9WrrysvgsCAACAFlVTYLJ8+fKsW7cuTz31VLZs2ZJnn302q1atqirHSZLVq1fnxRdfzNatW9Pd3Z0XXngha9euTZKMj4/nG9/4Ri655JL87d/+bdWJlUOHDmXBggXp7e3N5ORknn766axbt+6M66rlSE0ze3/iRAnOFUsXtM19AwAAwHyruSTnkUceydatW/PQQw9l6dKlM2U227Zty8aNG3PDDTfkjjvuyN69ezM4OJju7u709/fn4YcfTpI8/fTT+dnPfpY/+IM/yIYNG5Ik1157bf7mb/4me/fuzfbt21MoFDIxMZGrr776pBMo7axypHB/rwk5AAAAcKEUDh06dHITEhrS9//vD/N/vnw4SfLE9ZfmxtU9dV4RAAAAtCbHFJpIaWRy5nV/b2cdVwIAAACtTWDSRCpLcvp6PDoAAAC4UOy6m0hppCIwccIEAAAALhiBSRMpDk+V5HxyUUe6OwpnuBoAAAA4VwKTJlEul2dOmOhfAgAAABeWwKRJ/O7YZMaO93zt178EAAAALig77yYxNGxCDgAAAMwXgUmTqGz42t8jMAEAAIALSWDSJKpGCvd6bAAAAHAh2Xk3iZKSHAAAAJg3ApMmURxWkgMAAADzRWDSJIqVPUyU5AAAAMAFZefdJCpLcvqcMAEAAIALSmDSJIaOnzC5dGFHFnQW6rwaAAAAaG0CkyZQLpdTOt7DRDkOAAAAXHh2303g/WOT+eh4RY6GrwAAAHDhCUyaQHHESGEAAACYTwKTJlAaNiEHAAAA5pPddxMYqghMTMgBAACAC09g0gRKSnIAAABgXglMmkDlCZP+Ho8MAAAALjS77yZQGqnsYeKECQAAAFxoApMmUBw+UZKjhwkAAABceAKTJlA8XpJzycJCFnUV6rwaAAAAaH0CkwZXLpdnSnL6nS4BAACAeSEwaXAffFTO6PEWJn36lwAAAMC8EJg0uOKICTkAAAAw3+zAG1xx2IQcAAAAmG8CkwZXOSFHYAIAAADzQ2DS4EpVJTkCEwAAAJgPApMGN1RRktPX63EBAADAfLADb3ClipKcASU5AAAAMC8EJg2uckpOn5IcAAAAmBcCkwY3PSXn4gWF9HQV6rwaAAAAaA8CkwZWLpdnpuRo+AoAAADzR2DSwD4cK2dkopzESGEAAACYTwKTBlYyIQcAAADqwi68gQ1VTMhRkgMAAADzR2DSwEoVE3KU5AAAAMD8EZg0sGJFSU5/j0cFAAAA88UuvIEVnTABAACAuhCYNLBSZQ8TgQkAAADMG4FJAxuqnJKjJAcAAADmjV14A5tu+npRdyGLuz0qAAAAmC924Q2seLwkp085DgAAAMwrgUmDOjw2maPj5SQm5AAAAMB8sxNvUFUjhZ0wAQAAgHklMGlQRRNyAAAAoG4EJg1quuFrYkIOAAAAzDc78QY1pCQHAAAA6kZg0qBKSnIAAACgbgQmDapYUZJjSg4AAADMLzvxBlU5JafPCRMAAACYVwKTBlUamSrJWdJVyNJujwkAAADmk514g5o+YdLX6xEBAADAfLMbb0BHxiZzeKycRMNXAAAAqAeBSQOqmpDTIzABAACA+SYwaUBVE3KcMAEAAIB5JzBpQKVhI4UBAACgnuzGG9DQyImSHCOFAQAAYP4JTBpQ1QkTgQkAAADMO4FJAyoqyQEAAIC6shtvQMWKkhwnTAAAAGD+CUwa0PQJk96uQpZ2F+q8GgAAAGg/ApMGND1WuL+nI4WCwAQAAADmm8CkwQyPT+bDj8pJTMgBAACAehGYNJjScEX/kh6BCQAAANSDwKTBTJfjJEl/r8cDAAAA9WBH3mCqRgoryQEAAIC6EJg0mOKwkcIAAABQbwKTBlOqLMnp8XgAAACgHuzIG8xQRUmOKTkAAABQHwKTBlMaMSUHAAAA6k1g0mCmm74u6kwuXlCo82oAAACgPQlMGsx0YNLf25lCQWACAAAA9VBzYPL666/nK1/5Sq655ppcf/31efXVV0+6plwu5zvf+U6uvfbaDA4O5qabbsobb7wx8/kLL7yQz3/+8/nc5z6XW2+9NUeOHJn57Ne//nXWr1+fa665Jps2bUqxWDzHW2s+I+PlHPqonEQ5DgAAANRTzYHJ9u3bc9ttt2X37t258847s23btpOuef7557Nr167s2LEju3btynXXXZf7778/SXLkyJFs27YtTzzxRF5++eX09/fnwQcfTDIVtNxxxx154IEHsnv37nz5y1/OPffcc5632DyqJuRo+AoAAAB1U1NgcvDgwezZsyebN29OkmzatCn79u3Lvn37Trr22LFjGR0dTblczuHDh3P55ZcnSf71X/81n/vc53LllVcmSf7qr/4q//zP/5wkefnll7Nw4cJs2LAhSfL1r389zz33XMbGxs79DptIqXJCjpHCAAAAUDddtVy8f//+DAwMpKtr6muFQiErVqzI22+/ndWrV89ct3HjxuzcuTNXXXVVlixZkoGBgTz33HNJkrfeeisrV66cuXbVqlUZGhrK5OTkSZ8tXbo0S5YsSbFYrHp/ttHR0Vpuo2G9+cGxmdefXFBumfsCAACARrBo0aKzvramwCTJSY1Iy+XySdfs2bMnr732Wl555ZVcdNFFue+++3L33Xfn0Ucf/di/Uevfn+3AgQOZmJg443WN7v8b6kqyIEnSPfx+3nrr3fouCAAAAFpEZ2dn1qxZc9bX1xSYXHHFFTlw4EDGx8fT1dWVcrmc/fv3Z8WKFVXXPfnkk9mwYUOWLVuWJLnlllty8803J0lWrlyZHTt2zFz75ptvZmBgIB0dHVm5cmXefPPNmc8OHz6cI0eOpL+//7Trmi73aXbH3juaZOpUyWeu+GRWDiyo74IAAACgTdUUmCxfvjzr1q3LU089lS1btuTZZ5/NqlWrqspxkmT16tV58cUXs3Xr1nR3d+eFF17I2rVrkyRf/OIX861vfSt79+7NlVdemX/4h3/In//5nydJPvvZz2Z0dDQ7duzIhg0b8uMf/zg33nhjuru7T7uuWo7UNLJ3x0ZmXq9a1pNFi05/3wAAAMCFUXNJziOPPJKtW7fmoYceytKlS2fKbLZt25aNGzfmhhtuyB133JG9e/dmcHAw3d3d6e/vz8MPP5xkqi/JD37wg2zZsiXj4+NZu3btzN/o6OjIY489lrvuuiujo6MZGBjIY489Noe329gqm76akgMAAAD1Uzh06NCZm4QwLwb/pZRXDo1nYWdSvPXy0/Z6AQAAAC4cs2sbSHFkMknS19MpLAEAAIA6Epg0iGMT5fzu2FRg0t+jHAcAAADqSWDSIEojlf1LPBYAAACoJzvzBlGsaPjap+ErAAAA1JXApEEUhydnXivJAQAAgPoSmDQIJTkAAADQOOzMG0RlSU6/khwAAACoK4FJg5geKZwoyQEAAIB6E5g0iOoTJh4LAAAA1JOdeYOYDky6O5JLF3osAAAAUE925g2idLwkp6+nM4VCoc6rAQAAgPYmMGkAH02U8+7oVGCiHAcAAADqz+68AbxTOVJYw1cAAACoO4FJA6iakGOkMAAAANSdwKQBVE/IEZgAAABAvQlMGkBlYNLX45EAAABAvdmdNwAlOQAAANBYBCYNoKQkBwAAABqKwKQBVPUwUZIDAAAAdWd33gCmS3K6CsknFnkkAAAAUG925w1g+oRJX09nOgqFOq8GAAAAEJjU2fhkOe+OTp0w6e/1OAAAAKAR2KHX2Tsjkykff92n4SsAAAA0BIFJnVU3fBWYAAAAQCMQmNRZcaRypLDHAQAAAI3ADr3OisOTM6/7leQAAABAQxCY1FnVCRMlOQAAANAQBCZ1VtnDpE9JDgAAADQEO/Q6K2n6CgAAAA1HYFJnxZGpHiadheSTizwOAAAAaAR26HU2XZJzWU9HOjsKdV4NAAAAkAhM6mp8spyDo1MnTEzIAQAAgMYhMKmjg6OTmSxPve7TvwQAAAAahsCkjiobvg6YkAMAAAANwy69joojFSOFnTABAACAhiEwqaPi8OTMaz1MAAAAoHEITOqoWFGS068kBwAAABqGXXodVQUmSnIAAACgYQhM6qg4oiQHAAAAGpHApI6mT5h0FJLlizwKAAAAaBR26XVUOj4lZ/mijnR2FOq8GgAAAGCawKROJibLeed4SY5yHAAAAGgsApM6eXd0MhPlqdf9PR4DAAAANBI79TopjlSOFHbCBAAAABqJwKROisMnJuT0CUwAAACgoQhM6qRUecKkR2ACAAAAjURgUifTI4WTpL/XYwAAAIBGYqdeJ5UlOU6YAAAAQGMRmNSJpq8AAADQuAQmdTJdklNIcpmxwgAAANBQ7NTrpHS8JGd5T0e6Ogp1Xg0AAABQSWBSB5Pl8syUnD79SwAAAKDhCEzq4L3RyYyXp173K8cBAACAhmO3XgfFkYoJORq+AgAAQMMRmNTBdMPXJOkTmAAAAEDDEZjUQWVgMtDrEQAAAECjsVuvg6oTJpq+AgAAQMMRmNRBSQ8TAAAAaGgCkzqoPGFiSg4AAAA0Hrv1efb6B+P59cGPZv77yFi5jqsBAAAAPo7AZB49/trRfP5fSlVjhQefeSdPvHa0jqsCAAAAZiscOnTIEYd58PoH4/n8v5Qy+TH/tzsKya//l76suahr/hcGAAAAnMQJk3ny+GtHUzjFZ4UkP9nrlAkAAAA0CoHJPHnzyEROdZSnfPxzAAAAoDEITObJqiWdpz1hsmqJ8cIAAADQKAQm8+Qv/sfi054wufXKxfO5HAAAAOA0BCbz5FMXd+Xv/nhZOgpJZyFV//67P16m4SsAAAA0EFNy5tkbH47nJ3uP5s0jE1m1pDO3XrlYWAIAAAANRmACAAAAMIuSHAAAAIBZBCYAAAAAswhMAAAAAGYRmAAAAADMIjABAAAAmEVgAgAAADBLV61feP311/PNb34z7733Xi6++OL86Ec/yqc//emqa37605/m7//+72f++8CBAxkcHMzjjz+eX/7yl/n2t78989m7776byy67LC+99FKSZNmyZVm7dm06OqaynO9973sZHBw8p5sDAAAAOBeFQ4cOlWv5wp/+6Z/mq1/9arZs2ZJnnnkmP/zhD/OLX/zitN8ZHBzMX//1X+fP/uzPTvps8+bNWb9+fbZt25ZkKjB5++23s2TJklqWBQAAADBnairJOXjwYPbs2ZPNmzcnSTZt2pR9+/Zl3759p/zO7t2788477+SGG2446bOhoaG89NJL+epXv1rjsgEAAAAunJpKcvbv35+BgYF0dU19rVAoZMWKFXn77bezevXqj/3OT37yk2zevDnd3d0nffbTn/40X/rSl7J8+fKq92+66aaMjY3luuuuyz333JPFixefdl2jo6O13AYAAADQhhYtWnTW19bcw6RQKFT9d7l86oqe4eHhPP300/n5z3/+sZ8/8cQTeeCBB6re+81vfpOVK1fm6NGjueuuu3LvvffmwQcfPO2aDhw4kImJibO8AwAAAKDddHZ2Zs2aNWd9fU2ByRVXXJEDBw5kfHw8XV1dKZfL2b9/f1asWPGx1z/zzDO56qqrTmoKmyS/+tWvMjIyki9+8YtV769cuTJJsnjx4tx+++3Zvn37Gdd1+eWX13IbAAAAAKdVU2CyfPnyrFu3Lk899VS2bNmSZ599NqtWrTplOc7jjz+eW2+99ZSffe1rX0tnZ+fMe4cOHcqCBQvS29ubycnJPP3001m3bt0Z11XLkRoAAACAM6l5Ss5rr72WrVu35ne/+12WLl2aRx99NJ/5zGeybdu2bNy4caa5629/+9ts2LAh//Vf/5WlS5dW/Y3Dhw/nM5/5THbu3Jnf+73fm3n/3//937N9+/YUCoVMTEzk6quvzne/+91ccskl53+nAAAAAGep5sAEAAAAoNXVNFYYAAAAoB0ITAAAAABmEZgAAAAAzCIwAQAAAJhFYAIAAAAwi8AEAAAAYBaBCQAAAMAsAhMAAACAWQQmAAAAALMITAAAAABmEZgAAAAAzCIwAQAAAJhFYAIAAAAwi8AEAAAAYBaBCQAAAMAsAhMAAACAWQQmAAAAALMITAAAAABm+f8B4h7Uf/OvouEAAAAASUVORK5CYII=",
      "text/plain": [
       "<Figure size 1200x800 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\u001b[1m115/115\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 2ms/step\n",
      "\u001b[1m25/25\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 4ms/step\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "<Figure size 1200x600 with 0 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 1600x800 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "<Figure size 1200x600 with 0 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 1600x800 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Training set accuracy:   0.967\n",
      "Test set accuracy:       0.962\n",
      "Training set\n",
      "Class 0: precision=0.918, recall=0.958, f1_score=0.937\n",
      "Class 1: precision=0.985, recall=0.970, f1_score=0.978\n",
      "Test set\n",
      "Class 0: precision=0.915, recall=0.925, f1_score=0.920\n",
      "Class 1: precision=0.977, recall=0.973, f1_score=0.975\n"
     ]
    }
   ],
   "source": [
    "history_model1_df = pd.DataFrame(history_model1.history)\n",
    "plot_measure(history_model1_df.loss, history_model1_df.val_loss, \"Loss\")\n",
    "plot_measure(history_model1_df.f1_score, history_model1_df.val_f1_score, \"F1 Score\")\n",
    "\n",
    "model1.load_weights(model_file)\n",
    "predictions_train_model1 = np.argmax(model1.predict(X_train), axis=-1)\n",
    "predictions_test_model1 = np.argmax(model1.predict(X_test), axis=-1)\n",
    "\n",
    "plot_confusion_matrix(y_train, predictions_train_model1)\n",
    "plot_confusion_matrix(y_test, predictions_test_model1)\n",
    "\n",
    "print_accuracy_scores(y_train, y_test, predictions_train_model1, predictions_test_model1)\n",
    "print_stats(y_train, predictions_train_model1, \"Training set\")\n",
    "print_stats(y_test, predictions_test_model1, \"Test set\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "In tutti i 3 tuning i filtri dei layer Conv2D e il numero di nodi dei layer Dense è stato lo stesso.\n",
    "\n",
    "Per cnn0 non è stato utilizzati la regolarizzazione e il dropout rate è stato 0.2.\n",
    "\n",
    "Per i 2 tuning di cnn1 si è ottenuto lo stesso risultato, dropout rate pari a 0.5 e L2 pari a 0.001. Nonostante si sia ottenuta la stessa rete i risultati differiscono di non poco e ciò dipende da Tensorflow. Impostare il seed non è sufficiente infatti andrebbe utilizzata l'opzione sperimentale tf.config.experimental.enable_op_determinism(), che non è ancora disponibile per il layer MaxPool2D."
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "ml-project",
   "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": "3.11.8"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}